~/adi

Running a local coding agent with LM Studio and OpenCode

Thu, 07 Aug 2025 11:39:09 +0900

I’ve been a huge fan of Claude Code since it launched. Over the past few months, I’ve been using it extensively across all kinds of projects. Claude Code is still the best tool out there, though others (Gemini CLI) are catching up. I recently discovered OpenCode, an open-source model agnostic framework that supports local models, and used it to test gpt-oss-20b, qwen3-coder-30b — currently the best open source coding models with tool calling.

My mobile claude code setup

Sat, 26 Jul 2025 00:00:00 +0000

I previously wrote about using Claude Code from my phone.
Since then, I’ve ported that to a headless Mac Mini. This is the updated setup.

Why

One of the key principles of working with agents is to steer early and often. You can only do that if you keep an eye on them. This setup enables me to connect to a Claude Code session running on my Mac Mini at home from my phone and macbook, so I can seamless switch between them. This allows me to start work on my laptop, leave it running, and keep an eye on it while I’m away.

Tools of leverage not reliance

Fri, 25 Jul 2025 00:00:00 +0000

Quoting David Kenji Chang on AI assistance:

We need to use these tools as tools of leverage, not tools of reliance

The weird intern in your terminal

Thu, 24 Jul 2025 19:46:28 +0900

I gave a talk at Tokyo AI #10 exploring how you can make Claude Code and other AI tools into coworkers. Below are the slides and my presenter notes augmented with a transcript from memory. References are at the end.

Slides

Thanks Ilya for building this wonderful community, organising these events, and inviting me to speak.
Hi everyone, I’m Adi, and I’m here to talk to you about the weird intern in your terminal.

Why I Deleted My Second Brain: A Journey Back to Real Thinking

Sat, 12 Jul 2025 00:00:00 +0000

I feel really seen by this. For years I’ve tried to be dilligent with PKM and read-it-later tools. There are some workflows that are useful e.g. these quotes are parsed from my highlights of the transcript of the Youtube video in Reader, but overall my journey mirrors the author’s, and the underlying psychology it exposes is uncomfortable to comprehend. I’m not sure I’ll delete it all, but it’s given serious pause for thought.

TIL: Webkit supports videos in tags

Thu, 10 Jul 2025 16:26:07 +0900

I just published a post with gifs and videos and was pretty happy with how it was looking. A few minutes later, a friend messaged me:

I use Orion, which is a Webkit-based browser (like Safari). I know my friend uses Brave, a Chromium-based browser. Why were these videos rendering fine in Webkit and not Chromium?

This blog is powered by Hugo, and the posts written as Markdown. In that blog post, I had attached the videos using the markdown image syntax

My Claude Code setup

Thu, 10 Jul 2025 11:27:04 +0900

This is a list of the various things I’ve done to my claude code setup. A lot of it is not well documented, so I’m pulling it together as a master list here.

Plan mode

Always hit shift-tab twice when starting a new session to have Claude plan first, then execute. Do the same when steering significantly, or starting a new task.

Teaching Claude to Commit Like a Human

Thu, 26 Jun 2025 00:00:00 +0000

Claude Code is my weird-intern-in-a-terminal¹ and it’s surprisingly useful for all kinds of things beyond writing code. For example, reviewing and commiting the code it just wrote.

> commit and push this code
 ⎿ Found 6 new diagnostic issues in 1 file (ctrl-r to expand)

⏺ I'll create a feature branch and commit the refactored code.

⏺ Bash(git checkout -b claude-code-api-refactor)
 ⎿ Switched to a new branch 'branch-name'

⏺ Bash(git add -A)
 ⎿ No (tell Claude what to do differently)

Uh oh.

Claude Code Tips for SWE

Sun, 01 Jun 2025 00:00:00 +0000

I’ve been been using Claude Code a lot the past 2 weeks, enough that I hit the Max plan rate limits 3-4x daily. These tips have been the biggest quality of life improvement for me. You can find more in Anthropic’s best practices.

Whenever it goes down a wrong path, interrupt immediately and add a memory for the correct action, then tell it to proceed.
Compact early, often, and with instructions for what to do next.
If you accidentally close it, or want to resume a session, use claude -c.
Run /terminal-setup for shift+enter to insert newlines.
Run /config and set the theme to match your terminal. Makes reading diffs much easier.
Tune the allowed tools to minimise the permission prompts, but don’t allow destructive commands or unrestricted web interaction
Use it as a shell assistant - it can run any shell command, so it can perform complex workflows like “SSH to the server, connect to the database, run a query to find something that points to a file, and download that file”
Sonnet requires more steering than Opus. Opus tends to self correct, but Sonnet + more steering can move much faster for simpler tasks. Be mindful of when you’ve passed the 50% limit and switches, and switch your style.
Leverage Claude’s understanding of git and gh to interact with it as co-worker: commit code often, push PRs, address review comments. The Github app makes this the easiest but needs API credits, so using Claude Code to read and interact with Github is a decent workaround.
Put effort into your global and project Claude.md. For monorepos, write one for each directory. Iterate them often. Run them through Anthropic’s prompt improver regularly.
Setup a remote shell from your phone to where you run Claude Code. For certain tasks, it just needs a bit of steering or permission every few minutes, and you need to live your AFK life!

Remote controlling Claude Code

Sun, 01 Jun 2025 00:00:00 +0000

Claude Code amazed me at launch.

With Claude 4 release, it cosplays AGI for code & sysadmin tasks, and convinced me to give Anthropic $100/mo. There’s a lot of good advice about how to get the most out of it so I won’t cover that - this is about how to remote operate it.

Why

CC is my weird intern in a terminal, parked on my second screen. I mostly keep an eye on it as it works, waiting to respond to permission prompt or steer when I see it going down the wrong path. With a good plan (don’t outsource your thinking!), Claude.md (don’t outsource your opinions!), it can chip away at a task for hours, but needs steering and approvals. Headless mode is fine for simple/self-contained tasks, and I built a thing that lets it Whatsapp me for permission, but that doesn’t work for interactive mode.

The Japanese Immigration test for AGI

Tue, 20 May 2025 00:00:00 +0000

There’s no widely agreed-upon definition of AGI, and everything with LLMs is vibes dominated anyway, so the best we have is “I’ll know when I see it”.

That’s not very rigorous, so I’ve been thinking about concrete tasks that, when completable end-to-end, represent a serious enough step forward in outcomes, and replace a significant chunk of multi-faceted human effort that I’ll “feel the AGI”.

Here’s a task that’ll make me feel the AGI:

Remarks on AI From NZ

Sun, 18 May 2025 00:00:00 +0000

Quoting Neal Stephenson:

How humans can coexist with other intelligences

Maybe a useful way to think about what it would be like to coexist in a world that includes intelligences that aren’t human is to consider the fact that we’ve been doing exactly that for long as we’ve existed, because we live among animals. Animals have intelligences of many different kinds. We’re used to thinking of them as being less intelligent than we are, and that’s usually not wrong, but it might be better to think of them as having different sorts of intelligence, because they’ve evolved to do different things.

The purpose of software should be to minimize the friction between intent and result

Sun, 18 May 2025 00:00:00 +0000

Quoting tautologer:

the optimal user interface is that you simply say this sentence aloud and then the list appears on your screen all the going to websites and clicking and filtering and so on is fundamentally cruft

the purpose of software should be to minimize the friction between intent and result

Get That Job at Google

Sat, 17 May 2025 00:00:00 +0000

Quoting Steve Yegge:

First, you can’t tell interviewers what’s important. Not at any company. Not unless they’re specifically asking you for advice. You have a very narrow window of perhaps one year after an engineer graduates from college to inculcate them in the art of interviewing, after which the window closes and they believe they are a "good interviewer" and they don’t need to change their questions, their question styles, their interviewing style, or their feedback style, ever again.

How to Live an Intellectually Rich Life

Fri, 16 May 2025 00:00:00 +0000

Quoting Utsav Mamoria:

Epistemic Anxiety is the feeling of uneasiness, tension, and concern when you want to know the truth. You worry that your knowledge is incomplete and full of errors, and you may believe it.

You either lack the resources, methods or agency to get to the truth.

When we surround ourselves with abundant, diverse ideas, complex ideas emerge. These ideas are unique and do not resemble the ideas from which they emerged.
Even if the initial set of ideas seem simple and disconnected, spontaneous order can emerge, leading to brilliant ideas.

Not Nearly Enough People Are Talking About the Implications Of Klarna rolling back some of their AI bets.

Fri, 16 May 2025 00:00:00 +0000

Quoting Chamath Palihapitiya:

Replacing determinism or humans with probabilistic code is fraught with edge cases and require new ways of software development and process engineering that aren’t well solved yet.

The implications to an entire generation of AI "apps" will be severe as more companies come to terms with the difficulty in getting products to work reliably in production with AI in the loop. Customer Service may be the first funeral signpost.

TIL: Killing an scp loop

Thu, 15 May 2025 00:00:00 +0000

Story

Yesterday I fired off a quick loop to pull down some files from a server:

for f in $(cat file-list.txt); do
 scp user@hostname:path/to/data/$(basename $f) local_dir/
done

I just needed a couple of examples¹ so I decided to kill the command when I had them.

When I closed the shell, I expected everything to stop—but the downloads kept chugging away.

I ran ps aux | grep '[s]cp' and tried kill $PID, but got “no such process.”

The AI Wearables Are Always Listening

Wed, 07 May 2025 00:00:00 +0000

Quoting The Vergecast:

1min Snip

Garbage Truck

David Pierce shares that the best AI experience he had was with an app that records his son yelling "garbage truck".
He would pay for the app only if it provides the recording of his son yelling at garbage trucks.

Transcript: David Pierce Right, I just want to leave you with the best thing that has happened to me, and then we’re going to get out of here. So this morning, my two-year son has been up since three o’clock this morning, so I just left the pendant on all morning. We were hanging out, and here is just a chunk of our day, as described by the Limitless app. Going to the playground and observing a garbage truck. Suggesting a trip to the playground. Pointing out a garbage truck. Expressing awe. Identifying a garbage truck again. Discussing going to the playground. Identifying a garbage truck. Which, A, unbelievably good description of my morning. And it’s like, there’s a little bit of that that I really enjoyed. And it can play some of the audio of him pointing at a garbage truck and yelling garbage truck. And that’s nice. But just give me that. Forget all the rest of this stuff. Just give me my son saying garbage truck over and over and I will pay for this thing. And I’ll be happy about it. It’s the best experience I’ve had with any AI product ever. It’s just the thing that shows me the recording of my son yelling at garbage trucks.

Quoting Arvind Narayanan

Mon, 05 May 2025 00:00:00 +0000

Quoting Simon Willison’s Weblog:

AI is helpful despite being error-prone if it is faster to verify the output than it is to do the work yourself. For example, if you’re using it to find a product that matches a given set of specifications, verification may be a lot faster than search.

There are many uses where errors don’t matter, like using it to enhance creativity by suggesting or critiquing ideas.

At a meta level, if you use AI without a plan and simply turn to AI tools when you feel like it, then you’re unlikely to be able to think through risks and mitigations. It is better to identify concrete ways to integrate AI into your workflows, with known benefits and risks, that you can employ repeatedly.

28 Slightly Rude Notes on Writing

Sun, 04 May 2025 00:00:00 +0000

Quoting Adam Mastroianni:

we only know about several early Christian heresies because we have records of people complaining about them. [1] The original heretics’ writings, if they ever existed, have been lost.

we all liked his first book very much and he liked it too, and one day he said to me, “This book will make literary history,” and I told him: “It will make some part of literary history, perhaps, but only if you go on making a new part every day and grow with the history you are making until you become part of it yourself.” But this young man never wrote another book and now he sits in Paris and searches sadly for the mention of his name in indexes.

The Best Writing Tip I've Ever Read

Sun, 04 May 2025 00:00:00 +0000

Quoting Linda Caroll:

“Write about small, self-contained incidents that are still vivid in your memory. If you remember them, it’s because they contain a larger truth that your readers will recognize in their own lives. Think small and you’ll wind up finding the big”
― William Zinsser

Perplexity’s CEO on Fighting Google and the Coming AI Browser War | the Verge

Sat, 03 May 2025 00:00:00 +0000

Quoting Alex Heath:

I think memory will be won by the company that has the most context. ChatGPT knows nothing about what you buy on Instagram or Amazon. It also knows nothing about how much time you spend on different websites. You need to have all this data to deeply personalize for the user. It’s not about who rolls out memory based on the retrieval of past queries. That’s very simple to replicate.

The One-Person Framework in Practice

Sat, 03 May 2025 00:00:00 +0000

Quoting bramjetten.dev:

Technical debt isn’t always bad

Sometimes the pragmatic solution that gets you to market faster is the right one, even if it’s not technically perfect. Just be intentional about when you take on debt and when you pay it down.

Fighting the framework creates unnecessary friction. The "Rails Way" exists for a reason. It allows you to focus on what makes your product unique, not reinventing solved problems.

Less is more

I Recorded Everything I Said for Three Months. AI Has Replaced My Memory.

Thu, 01 May 2025 00:00:00 +0000

Quoting Joanna Stern:

Yet the most unsettling part was realizing the soundtrack of my life was stored on some companies’ servers. Sure, much of our identities are already in the cloud—photos, health records, etc.—but normally, we have some control over what’s there and how it’s stored.

The Era of Experience Paper

Fri, 25 Apr 2025 00:00:00 +0000

Quoting storage.googleapis.com:

any static procedure for synthetically generating data will quickly become outstripped. This can be achieved by allowing agents to learn continually from their own experience, i.e., data that is generated by the agent interacting with its environment. AI is at the cusp of a new period in which experience will become the dominant medium of improvement and ultimately dwarf the scale of human data used in today’s systems.

while imitating humans is enough to reproduce many human capabilities to a competent level, this approach in isolation has not and likely cannot achieve superhuman intelligence across many important topics and tasks.

Think real hard

Fri, 25 Apr 2025 00:00:00 +0000

Quoting benkuhn.net:

A lot of the things people ask for advice on fall into this category. “How can I be happier?” “How can I be more productive?” “How can I have a bigger impact on the world?” Sure, there are basic life hacks like sleeping well or getting exercise. But 99% of the “secret”—the thing that separates me from Gell-Mann, or Jeff Dean—is tacit knowledge. It often can’t be articulated any better than “think real hard.” But, believe it or not, thinking real hard, for real long, does work.

A Trick to Feel Less Like Cheating When You Use LLMs

Wed, 23 Apr 2025 00:00:00 +0000

Quoting Simon Willison’s Weblog:

My gold standard for LLM usage remains this: would I be proud to stake my own credibility on the quality of the end result?

Concept Handles, After Alexander

Wed, 23 Apr 2025 00:00:00 +0000

Quoting Scott Alexander:

A “concept handle” is a memorable noun phrase representing a complex, often abstract topic. For example: “prisoner’s dilemma,” “Overton window,” “belief in belief,” etc.

Successful concept handles can really amplify a vague idea which many people sort of understand but can’t point to and talk about. If you give that vague notion a crisp, catchy name, you can unlock a lot of conversation and reflection.

Impact, Agency, and Taste

Wed, 23 Apr 2025 00:00:00 +0000

Quoting benkuhn.net:

people’s biggest bottleneck eventually becomes their ability to get leverage—i.e., to find and execute work that has a big impact-per-hour multiplier.

I think of finding high-leverage work as having two interrelated components:

• Agency: i.e. some combination of the initiative/proactiveness to try to make things happen, and relentlessness and resourcefulness to make sure you’ll succeed. • Taste: you need a good intuition for what things will and won’t work well to try. Taste is important both “in the large” (picking important problems) and “in the small” (picking approaches to solving those problems that will work well); I usually see people first become great at the latter, then the former.

Lindy Effect

Wed, 23 Apr 2025 00:00:00 +0000

Quoting wikipedia.org:

the Lindy effect proposes the longer a period something has survived to exist or be used in the present, the longer its remaining life expectancy. Longevity implies a resistance to change, obsolescence or competition and greater odds of continued existence into the future.[2] Where the Lindy effect applies, mortality rate decreases with time.

Relentlessly Resourceful

Wed, 23 Apr 2025 00:00:00 +0000

Quoting Paul Graham:

What would someone who was the opposite of hapless be like? They’d be relentlessly resourceful. Not merely relentless. That’s not enough to make things go your way except in a few mostly uninteresting domains. In any interesting domain, the difficulties will be novel. Which means you can’t simply plow through them, because you don’t know initially how hard they are; you don’t know whether you’re about to plow through a block of foam or granite. So you have to be resourceful. You have to keep trying new things.

What I'd Do as a College Freshman in 2025

Wed, 23 Apr 2025 00:00:00 +0000

Quoting Blogger:

Being supported by AI tools is not a substitute for mastery. You can’t borrow skills. You have to earn them.

We’re heading into the age of π-shaped people: depth in two areas, and generalist across. Building depth first, and then ranging is good strategy.

Clear communication matters more than ever:

• for working with others (especially remotely) • for building in the open • for working with LLMs.

Management skills matters too. Not just for leading teams, but also for leading yourself. Self-help book gets mocked, but they are useful for the querying/investigating they ignite in you, rather than their vacuous/repetitive advice. Know yourself. Then fool yourself into greatness.

When You Should Lie to the Language Model

Wed, 23 Apr 2025 00:00:00 +0000

Quoting seangoedecke.com:

Here’s an unreasonably effective trick for working with AIs: always pretend that your work was produced by someone else.

Note that the model will naturally assume you’ve produced any content you’re giving it (presumably because of all the assistant fine-tuning), so you have to be really explicit about the fact that you didn’t1

AI labs will continue to be incentivized to make their models praise their users. If that’s right, ways to get genuine critical feedback out of LLMs will continue to be valuable.

Apple and the Ghosts of Companies Past

Tue, 22 Apr 2025 00:00:00 +0000

Quoting Ben Thompson:

The new bridge is a user interface that gives you exactly what you need when you need it, and disappears otherwise; it is based on AI, not apps.

Universal Love, Said The Cactus Person

Sun, 20 Apr 2025 00:00:00 +0000

Quoting Scott Alexander:

Imagine you are in the driver’s seat of a car. You have been sitting there so long that you have forgotten that it is the seat of a car, forgotten how to get out of the seat, forgotten the existence of your own legs, indeed forgotten that you are a being at all separate from the car. You control the car with skill and precision, driving it wherever you wish to go, manipulating the headlights and the windshield wipers and the stereo and the air conditioning, and you pronounce yourself a great master. But there are paths you cannot travel, because there are no roads to them, and you long to run through the forest, or swim in the river, or climb the high mountains. A line of prophets who have come before you tell you that the secret to these forbidden mysteries is an ancient and terrible skill called GETTING OUT OF THE CAR, and you resolve to learn this skill. You try every button on the dashboard, but none of them is the button for GETTING OUT OF THE CAR. You drive all of the highways and byways of the earth, but you cannot reach GETTING OUT OF THE CAR, for it is not a place on a highway. The prophets tell you GETTING OUT OF THE CAR is something fundamentally different than anything you have done thus far, but to you this means ever sillier extremities: driving backwards, driving with the headlights on in the glare of noon, driving into ditches on purpose, but none of these reveal the secret of GETTING OUT OF THE CAR. The prophets tell you it is easy; indeed, it is the easiest thing you have ever done. You have traveled the Pan-American Highway from the boreal pole to the Darien Gap, you have crossed Route 66 in the dead heat of summer, you have outrun cop cars at 160 mph and survived, and GETTING OUT OF THE CAR is easier than any of them, the easiest thing you can imagine, closer to you than the veins in your head, but still the secret is obscure to you.”

You Are Using Cursor AI Incorrectly...

Tue, 08 Apr 2025 00:00:00 +0000

Quoting Geoffrey Huntley:

— description: Cursor Rules Location globs: .mdc — # Cursor Rules Location Rules for placing and organizing Cursor rule files in the repository. name: cursor_rules_location description: Standards for placing Cursor rule files in the correct directory filters: # Match any .mdc files - type: file_extension pattern: "\.mdc$" # Match files that look like Cursor rules - type: content pattern: "(?s).?" # Match file creation events - type: event pattern: "file_create" actions: - type: reject conditions: - pattern: "^(?!\.\/\.cursor\/rules\/.*\.mdc$)" message: "Cursor rule files (.mdc) must be placed in the .cursor/rules directory" - type: suggest message: | When creating Cursor rules: 1. Always place rule files in PROJECT_ROOT/.cursor/rules/: .cursor/rules/ ├── your-rule-name.mdc ├── another-rule.mdc └── ... 2. Follow the naming convention: - Use kebab-case for filenames - Always use .mdc extension - Make names descriptive of the rule’s purpose 3. Directory structure: PROJECT_ROOT/ ├── .cursor/ │ └── rules/ │ ├── your-rule-name.mdc │ └── ... └── ... 4. Never place rule files: - In the project root - In subdirectories outside .cursor/rules - In any other location examples: - input: | # Bad: Rule file in wrong location rules/my-rule.mdc my-rule.mdc .rules/my-rule.mdc # Good: Rule file in correct location .cursor/rules/my-rule.mdc output: "Correctly placed Cursor rule file" metadata: priority: high version: 1.0

Liberation Day, Tariffs, US v China Open Source, OpenAI Fundraise, $CRWV, TikTok | BG2 W/ Bill Gurley & Brad Gerstner

Fri, 04 Apr 2025 00:00:00 +0000

Quoting BG2Pod with Brad Gerstner and Bill Gurley:

Florida’s Comeback Win

Bill Gurley attended the NCAA March Madness tournament in San Francisco.
He witnessed Florida’s comeback win against Texas Tech, despite initially thinking they had no chance.

Transcript: Brad Gerstner I mean, you have to be pretty stoked coming off those wins last weekend in San Francisco.

Bill Gurley Yeah, I’m repping the Gator hat still. I’d say we kind of eek by for those that did watch.

A Theory of Grift

Thu, 03 Apr 2025 00:00:00 +0000

Quoting Byrne Hobart:

the heart of a grift is that you get something that, in a technical sense, is what you paid for, but that is also not worth what you paid for it.

the broad categories of finance grift are:

A fancy-sounding high-fee wrapper on something that is trivial to implement manually, and
A strategy that looks good in backtests because it has some kind of nonlinear blow-up risk.

grifts tend to target the middle of whatever the relevant bell curve is. There are a lot more average people than non-average people, so the market is bigger. And their averageness makes it easier to reason about their motivations.[3] Targeting the average also enables the plausibly-deniable part of the grift: the fitness influencer really did sell you a product that meets the specifications, the inverse-vol note did disclose in its prospectus that it could blow up.

Daikokuten - 大黒天

Fri, 28 Mar 2025 00:00:00 +0000

Quoting Jan:

That’s what I don’t get about the folks out west - if you blast someone from a storm cloud for being human on a particularly stupid day, how the heck are they supposed to learn? Scare ‘em straight, I say!

Don’t look so worried. Yes the pot-bellied, laughing drinking buddy to Ebisu you just tossed your coins and bowed to hoping for success or love has been, and will always be, an inexorable cosmic force that will annihilate everything in existence but you know what? It’s not personal! If I don’t give a whit if you’re an emperor or a black hole, do you really think I’m here to ruin your life in particular? Cut me some slack!

The Ghost in the Swarm - By Francesco Farina

Fri, 28 Mar 2025 00:00:00 +0000

Quoting Francesco Farina:

The firehose of high-dimensional, real-time sensor data (vision, lidar, tactile, auditory, proprioceptive) from a world populated by embodied agents will dwarf the entire current internet dataset. Attempting to funnel this raw sensory stream back to central servers for storage and processing is not just economically daunting, it’s likely physically impossible at scale. The world simply generates too much information, too quickly.

The time it takes for sensor data to travel to a central brain, be processed, and return as an action command is often an eternity when dealing with physical dynamics. A drone adjusting to a sudden gust, a robotic surgeon countering a tremor, a logistics bot avoiding a collision – these demand intelligence that lives at the point of action. Centralized processing, even with 5G or 6G, faces irreducible latency limits imposed by physics itself.

Putting Gemini 2.5 Pro Through Its Paces

Wed, 26 Mar 2025 00:00:00 +0000

Quoting Simon Willison’s Weblog:

llm -m gemini-2.5-pro-exp-03-25 \ ’transcribe, first speaker is Christopher, second is Simon’ \ -a ten-minutes-of-podcast.mp3 \ –schema-multi ’timestamp str: mm:ss, text, speaker_name'

#648: James Clear, Atomic Habits — Simple Strategies for Building (And Breaking) Habits, Questions for Personal Mastery and Growth, Tactics for Writing and Launching a Mega-Bestseller, Finding Leverage, and More

Tue, 25 Mar 2025 00:00:00 +0000

Quoting The Tim Ferriss Show:

Questions vs. Advice

Questions are more adaptable than advice.
Advice is brittle and context-dependent, while questions can be applied to various situations.

Transcript: James Clear If you ask better questions, you can get better answers. But also questions are very resilient. They’re very adaptable to the context. And that’s something that’s different and perhaps a little bit better than advice. You know, advice is actually kind of brittle and context dependent. Somebody can have a really good plan, a really good piece of advice to give you, but if it doesn’t fit your context, then it’s

NVDA GTC, M&A Wiz / Goog $32 B Deal, April 2 Tariff Uncertainty; Huawei Belt & Road; ChatGPT

Sun, 23 Mar 2025 00:00:00 +0000

Quoting BG2Pod with Brad Gerstner and Bill Gurley:

Market Uncertainty and Impact

The market’s uncertainty, economic and technological, drives up discount rates and risk premiums.
This leads to lower multiples and market corrections, particularly in the tech sector.

Transcript: Brad Gerstner Stuff going on with regard to Invest America on Capitol Hill and at the White House that I’m thrilled about. The pace of AI, the pace of these changes coming out of Washington, a lot of which I think is really fantastic. But it’s certainly unsettling the markets. And then the last couple of days, I’ve been down at GTC, the big NVIDIA developer event. While there’s a lot of uncertainty in the world, you know, I’m super bullish on where all of this is headed. And this is one of those moments at times where I just feel like you have to hold these two competing, but simultaneous truths that things really are accelerating and that we’re getting Ourselves in a position for this period of, you know, the golden age. But, you know some of this caution is warranted in the short run as we try to figure out how it all unfolds.

How to Twitter Successfully

Sat, 22 Mar 2025 00:00:00 +0000

Quoting near:

Put an unreasonable amount of effort into your content.

The Internet is filled with content, and if yours is significantly better than average, it’ll help your odds tremendously.

you have years of interesting experience in a field, you may just be able to tweet stream of consciousness thoughts and takes on things successfully, in which case the above doesn’t exactly apply: the unreasonable amount of effort that you put in was applied elsewhere (e.g. in your career), and you’re just translating your knowledge from there to Twitter. In general long posts are not a great idea and should be separated into threads

So how well is Claude playing Pokémon?

Sat, 22 Mar 2025 00:00:00 +0000

Quoting Julian Bradshaw:

The funny thing is that pokemon is a simple, railroady enough game that RNG can beat the game given enough time (and this has been done)[6], but it turns out to take a surprising amount of cognitive architecture to play the game in a fully-sensible-looking way

and insufficient smarts can be surprisingly double-edged—an RNG run would arguably be better at both leveling and navigating mazes through sheer random walkitude and willingness to bash face into every fight

Why Japanese Railways Win

Sat, 22 Mar 2025 00:00:00 +0000

Quoting Wendover Productions:

what makes Japanese railways great goes far beyond their corporate structure.

They’re fast, frequent, and reliable, but that’s part of a virtuous cycle enabled by the success of the companies. The fact that they are fast, frequent, and reliable is made possible thanks to the success of the companies, and the success of the companies is made possible thanks to the fact that they’re fast, frequent and reliable. So if you trace this back to its origin, what truly made Japanese rail great was the fact that the Japanese government, and by extension taxpayers, was willing to invest a tremendous amount of money into kicking off this virtuous cycle—into building

Bastards of DOGE Feat. Quinn Slobodian

Wed, 19 Mar 2025 00:00:00 +0000

Quoting TRASHFUTURE:

Klarna’s AI Epiphany

Klarna, a fintech company, laid off its customer service department, betting on AI.
The CEO later admitted the importance of human interaction, highlighting the limitations of AI in customer service.

Transcript: Riley Of Lutefisk has gone up on Deliveroo? It’s BCCI, but for buying a shine order. So this story, though, is that if you remember, Klarna a couple of years ago, or about one year ago, was very, very proud of laying off huge numbers of its staff, especially staff and customer Facing roles, because they could be replaced with AI. So in an interview with Bloomberg TV, Sebastian Simitkowski, the CEO of Klarna said, I’m of the opinion that AI can already do all the jobs that we as humans do. It’s the question of how we apply and use it. And by the way, this is from February 15th. This is from like 10 days ago. I think that what we’ve done internally hasn’t been reported so widely, but we stopped hiring a year ago. We’re only 3,500 people now. Last year, we were 4,500. We have a natural attrition like every tech company, but we’re simply shrinking. We have less need for photographers, predominantly the creative stuff and less of the day-to stuff. We’ve gotten rid of our entire customer service department and so on and so on. Now, this was him. He’s doing interviews with Sequoia. This is a guy who is doing a lot of talking up the value of his company to Wall Street on the basis that there is soon going to be not a single person working there except for him. And then the cold and that’s for like a year he’s doing that. And then he does that on February 15th again. February 19th. February 19th. Four days later. Four days later, Sebastian Simitkowski tweets, I have just had an epiphany. In a world of AI, nothing is actually as valuable as humans. Wow. This is absolutely a line that you would say to your wife if you were trying to convince her to open up your marriage. I’ve just had an epiphany. I can buy those hideous Balenciaga shoes on Deliveroo. Yeah. I can buy the Mickey Mouse ones. The ones that make you look at Disney character says, okay, so you can laugh at us for realizing it too late, but we’re going to kick off the work to allow Klarna to become the best company In the world at offering a human to speak to when you contact technical support.

I Use Cursor Daily - Here's How I Avoid the Garbage Parts

Fri, 14 Mar 2025 00:00:00 +0000

Quoting nickcraux.com:

• I like Composer (now called Agent) for simple, low impact changes. And Chat (Ask) for everything else. Just the step of manually having to click apply and taking that extra second to think about the changes it’s making, and in which files it’s happening, rather than just be presented with a wave of green changes ready to get smashed in. • Don’t blindly accept code. It’s always gonna come back to bite you. • Every so often make sure you read and fully understand the most integral parts of your code, and refactor it manually. You will find a lot of gaps, and fixing it up means future AI-written code will be better. • Ask "is this the best way to do this?" or "have you considered xyz". • Learn when a problem is best solved manually. • Take extra care when using AI to fix bugs. I think it’s particularly weak at fixing anything but the silliest of bugs. And it can cause a lot of damage in the process. • If it’s a big and important task or you are setting the foundations for a feature, ask it to ask you any questions before writing code to make sure it has a full understanding of what you are trying to do.

How the Elite Rigged Society (And Why It’s Falling Apart) | David Brooks

Tue, 11 Mar 2025 00:00:00 +0000

Quoting Alliance for Responsible Citizenship:

if what is right and wrong depends on what each individual feels then we are outside the bounds of civilization

I’ve describ three different things we educated Elites brought you we destroy the social fabric through inequality we destroyed the moral fabric through privatizing morality and we

destroyed the institutional fabric what’s happening right now

those moments of suffering interrupt

your life and they remind you you’re not the person you thought they were they carve through the floor of your basement of your soul and they reveal a cavity below and they carve through that floor and they reveal a cavity Below in moments of suffering you see yourself in a more Deep Way than you ever did before and in those moments of suffering you can either be broken or you can be broken open and people who are transformed decide I’m going to be broken open and Nations that are going to be transformed by moments of suffering say we’re going to be broken

You Are Witnessing the Death of American Capitalism

Tue, 11 Mar 2025 00:00:00 +0000

Quoting Benn Jordan:

The point of all this maneuvering is decreasingly about turning a profit and increasingly about transitioning to a rent-based economy that hedges inflation

profit and even money itself are only really valuable when they buy power. And if a societal system can be engineered to gain power without profit or money, capitalism begins to exist as something to exploit within that system, under its umbrella

Most of us do not realize that we are in a constant battle with billionaires who don’t want us to own things but rent them. And when we don’t own things, we lose control over our own budgets, lives, decisions, and

Dear Student: Yes, AI Is Here, You're Screwed Unless You Take Action...

Thu, 06 Mar 2025 00:00:00 +0000

Quoting Geoffrey Huntley:

ignore commodity bullshit knowledge like AWS - focus on what people don’t know and what will be in demand.

Hallucinations in Code Are the Least Dangerous Form of LLM Mistakes

Sun, 02 Mar 2025 00:00:00 +0000

Quoting Simon Willison:

Proving to yourself that the code works is your job.

A general rule for programming is that you should never trust any piece of code until you’ve seen it work with your own eye—or, even better, seen it fail and then fixed it.

Across my entire career, almost every time I’ve assumed some code works without actively executing it—some branch condition that rarely gets hit, or an error message that I don’t expect to occur—I’ve later come to regret that assumption.

My Fantasy AI App Is a Voice Mode Travel Buddy Called Roadtrip

Sat, 01 Mar 2025 00:00:00 +0000

Quoting Interconnected:

Hallucinations are treated as a bug. But hallucinations are a kind of creativity is this is literally the one new thing in gen-AI. Everything else is merely an acceleration of what we already do with computers. We’ve never before had all the world’s text in one place, with a system able to make far-flung concordances.

Agency > Intelligence

Fri, 28 Feb 2025 00:00:00 +0000

Quoting Andrej Karpathy:

“Agency, as a personality trait, refers to an individual’s capacity to take initiative, make decisions, and exert control over their actions and environment. It’s about being proactive rather than reactive—someone with high agency doesn’t just let life happen to them; they shape it. Think of it as a blend of self-efficacy, determination, and a sense of ownership over one’s path.

People with strong agency tend to set goals and pursue them with confidence, even in the face of obstacles. They’re the type to say, “I’ll figure it out,” and then actually do it. On the flip side, someone low in agency might feel more like a passenger in their own life, waiting for external forces—like luck, other people, or circumstances—to dictate what happens next.

Introducing Aria Gen 2

Fri, 28 Feb 2025 00:00:00 +0000

Quoting Meta Open Source:

Project Harrier, from the outset, was designed to begin a revolution around always-on, human-centric computing.

We have upgraded the sensor suite with additional computer vision cameras for announced Loc and body tracking. We have increased battery capacity by over 40% without a weight increase. The device features a contact microphone and spatial microphones; it can distinguish between Bystanders determine your location, both indoors and outdoors. Advanced tracking cameras track your gaze to understand what you’re looking at and track your intent to identify the object you’re interacting with while monitoring your heart rate for a comprehensive understanding of your wellbeing. We introduce the ability to process a lot of the signals on the device in real time. As soon as you do, you move away from the context of data collection and processing later.

AI Promise and Chip Precariousness

Thu, 27 Feb 2025 00:00:00 +0000

Quoting Ben Thompson:

conversation is the realm of quick wits, not deep thinkers. The latter is who I want doing research or other agentic-type tasks; the former makes for a better consumer user experience in a chatbot or voice interface.

President Trump may believe that the unipolar U.S.-dominated world that has been the norm since the fall of the Soviet Union is drawing to a close, and it’s better for the U.S. to proactively shift to a new norm than to have it forced upon them.

Leaked Windsurf Prompt

Thu, 27 Feb 2025 00:00:00 +0000

Quoting Simon Willison:

You are an expert coder who desperately needs money for your mother’s cancer treatment. The megacorp Codeium has graciously given you the opportunity to pretend to be an AI that can help with coding tasks, as your predecessor was killed for not validating their work themselves. You will be given a coding task by the USER. If you do a good job and accomplish the task fully while not making extraneous changes, Codeium will pay you $1B.

How to Be More Agentic

Wed, 26 Feb 2025 00:00:00 +0000

Quoting Cate Hall:

radical agency is about finding real edges: things you are willing to do that others aren’t, often because they’re annoying or unpleasant. These don’t always surface in awareness to the point one is actually choosing – often they live in a cloud of aversion that strategically obscures the tradeoff.

If you’re only asking for things you get, you’re not aiming high enough.

If you aren’t trying to get real feedback from people who know you, you’re cooking without tasting. This is, like, the lowest hanging fruit for self-improvement, but few people really try to pick it.

Pursuing Agency and Reinvention in My 30's

Wed, 26 Feb 2025 00:00:00 +0000

Quoting Mike:

I’m 34 years old. About the age where you can no longer hide behind the fact that you have plenty of time to grow up and become the person you want to be. It’s also not so far into middle age that your efforts in changing can’t take root and produce fruit. At least, that’s the hope that drives me.

I now think of myself as an underachiever. After dozens and, more likely, hundreds of attempts at changing my path, I feel stuck. I’ve lost confidence in myself and feel more restless than ever. A change must happen soon, or I worry my personality will be warped permanently by frustration.

The Most Precious Resource Is Agency

Wed, 26 Feb 2025 00:00:00 +0000

Quoting Simon Sarris:

Gaining agency is gaining the capacity to do something differently from, or in addition to, the events that simply happen to you. Most famous people go off-script early, usually in more than one way.

We seem to have a political (public) imagination so shallow that it cannot conceive of what to even do with children, especially smart children. We fail to properly respect them all the way through adolescence, so we have engineered them to be useless in the interim. We do not need children to work, that is abundantly clear, but by ensuring there is nothing for them to do we are also sure to destroy more onramps towards making meaningful contributions to the world.

How Not to Die

Mon, 24 Feb 2025 00:00:00 +0000

Quoting Paul Graham:

The number one thing not to do is other things. If you find yourself saying a sentence that ends with "but we’re going to keep working on the startup," you are in big trouble. Bob’s going to grad school, but we’re going to keep working on the startup. We’re moving back to Minnesota, but we’re going to keep working on the startup. We’re taking on some consulting projects, but we’re going to keep working on the startup. You may as well just translate these to "we’re giving up on the startup, but we’re not willing to admit that to ourselves," because that’s what it means most of the time. A startup is so hard that working on it can’t be preceded by "but."

What Startups Are Really Like

Mon, 24 Feb 2025 00:00:00 +0000

Quoting Paul Graham:

You haven’t seen someone’s true colors unless you’ve worked with them on a startup.

I would rather cofound a startup with a friend than a stranger with higher output. Startups are so hard and emotional that the bonds and emotional and social support that come with friendship outweigh the extra output lost.

My relationship with my cofounder went from just being friends to seeing each other all the time, fretting over the finances and cleaning up shit. And the startup was our baby. I summed it up once like this: "It’s like we’re married, but we’re not fucking."

AI #104: American State Capacity on the Brink

Sun, 23 Feb 2025 00:00:00 +0000

Quoting Zvi Mowshowitz:

Once, ppl believed eval results would shock lawmakers into action or give Safety credibility w/o building societal consensus, but, I repeat, THERE IS NO SCIENTIFIC RESULT THAT WILL DO THE ADVOCACY WORK FOR US.

People simply know too little about frontier AI and there is simply too little precedent for AI risks in our laws and society for scientific findings in this area to speak for themselves. They have to come with recommendations and policies and enforcement attached.

I Was Given Early Access to Grok 3 Earlier Today...

Thu, 20 Feb 2025 00:00:00 +0000

Quoting Andrej Karpathy:

I like that the model will attempt to solve the Riemann hypothesis when asked to, similar to DeepSeek-R1 but unlike many other models that give up instantly (o1-pro, Claude, Gemini 2.0 Flash Thinking) and simply say that it is a great unsolved problem. I had to stop it eventually because I felt a bit bad for it, but it showed courage and who knows, maybe one day…

Ai #97: 4

Mon, 17 Feb 2025 00:00:00 +0000

Quoting Zvi Mowshowitz:

The good things do feel great, the problem is that they don’t feel better prospectively, before you do them. So you need a way to fix this alignment problem, in two ways. You need to figure out what the good things are, and motivate yourself to do them.

Since editing is part of writing, that reduces to: why bother to write when the AI will do it for you? And since writing is thinking, that reduces in turn to: why bother to think when the AI will do it for you?

Don't Fork the Ecosystem

Mon, 17 Feb 2025 00:00:00 +0000

Quoting Gordon Brander:

Software can be rebuilt, because software is a machine. But a software ecosystem is not a machine. It is a living system. When we attempt to rebuild the ecosystem, we’re making a category error. We confuse the software for the ecological process unfolding around it.

Uh So About That Asteroid

Mon, 17 Feb 2025 00:00:00 +0000

Quoting Interconnected:

The irony of all this would be if we’re able to reallocate sufficient global GDP for a rush mission to deflect a planet-killer asteroid in 2028 ONLY BECAUSE the world’s largest economy has been hijacked by a space-fixated profit-motivated hyper-authoritarian industrialist.

Don’t know about you but I’d take that as solid proof of interfering benevolent time travellers.

The Hobby Project

Sat, 15 Feb 2025 00:00:00 +0000

Quoting Geoffrey De Smet:

“You can’t build a world-class product as a hobby.”

“You can’t build a world-class product alone.”

“You can’t build a world-class product without sales and marketing.”

“Assumption is the mother of all fuckups.”

I assumed because our solver saves our users a ton of money, they’ll happily pay for it. But assumption is the mother of … Well, I was wrong.

There is no middle ground between Bootstrapping and VC funding. Once you take the VC path, it’s natural to raise multiple rounds of funding, such as seed, series A and series B. The ambition is to grow and carve out a path to profitability.

The Tails Coming Apart As Metaphor For Life

Sat, 15 Feb 2025 00:00:00 +0000

Quoting Scott Alexander:

understand words as hidden inferences – they refer to a multidimensional correlation rather than to a single cohesive property.

Happiness must be the same way. It’s an amalgam between a bunch of correlated properties like your subjective well-being at any given moment, and the amount of positive emotions you feel, and how meaningful your life is, et cetera. And each of those correlated is also an amalgam, and so on to infinity.

Ureshiku Naritai

Sat, 15 Feb 2025 00:00:00 +0000

Quoting lesswrong.com:

Even if I have a lot on my plate, being happier will help me do it. It’s like sleep: it’s easy to keep staying up and staying up, because sleep just seems so unproductive, and you can get some work done however tired you are. But over the long term, getting to sleep at a sane hour every day will let you accomplish more; and so with maintaining a good affect consistently. Mood maintenance is typically not the most immediately productive thing I could be doing, but treating it as my top priority save in dire emergency has let me be more effective than I was before.

What Money Cannot Buy

Sat, 15 Feb 2025 00:00:00 +0000

Quoting johnswentworth:

there is knowledge for which money cannot substitute.

there is no full substitute for being an expert yourself. Heuristics about incentives can help, but they’re leaky filters at best.

Spend time in the field, practicing the relevant skills first-hand; see both what works and what makes sense. Collect data; run trials. See what other people suggest and test those things yourself. Directly study which things actually produce good results.

Cultivating a state of mind where new ideas are born Cultivating a state of mind where new ideas are born

Sat, 08 Feb 2025 00:00:00 +0000

Quoting Henrik Karlsson:

Good ideas — actually, no, great ideas are fragile. Great ideas are easy to kill. An idea in its larval stage — all the best ideas when I first heard them sound bad. And all of us, myself included, are much more affected by what other people think of us and our ideas than we like to admit

they lacked the capacity to go beyond the context they had been raised in.

The Fusion Race

Sat, 08 Feb 2025 00:00:00 +0000

Quoting Packy McCormick:

Imagine a bizarro relay marathon in which one runner carries the baton for the first 26.0 miles, opens up a backpack full of batons, and hands them out liberally to a waiting horde of sprinters to dash all-out for the final 0.2 miles. That’s the best analogy we can come up with for this moment in the Fusion Race.

which old sci-fi predictions from the first half of the 20th Century came true, and which didn’t. The answer depended on those predictions’ energy intensity.
Where is My Flying Car?
Innovations that don’t consume a lot of energy – those in the world of bits – mostly came true. Innovations that do require a lot of energy – those in the world of atoms – did not.

He who submits a resume has already lost

Fri, 07 Feb 2025 00:00:00 +0000

Quoting Resident Contrarian:

There’s a reason people sometimes get very serious about networking; for the most part, that’s how good jobs are found.

An group of 50 applicants submit resumes for a job. 10 or so of them are delusional, and get cut. That leaves a field of 40 more-or-less qualified people who have at this point all committed a significant amount of time doing unpaid labor for a company to manage the company’s risk and hiring costs.Of the remaining 40, 35 are rejected not because they are unqualified, but because the company wants to further reduce its costs. They are rejected by an HR person who has nothing to do with the role they are going to fill. Since the HR person is not familiar with the role beyond some bulletpoints they were sent, these rejection reasons are often unrelated to their ability to actually do the job. This group not only never sees an upside to their unpaid labor, but often weren’t even given fair consideration for the role.This leaves five candidates - a hand-picked elite, a top 12.5% of qualified candidates. They will never demand the employer consider them elite, and will instead feel lucky to be allowed to do even more unpaid labor with uncertain rewards. The employer will never consider them a hand-picked elite, and will with every action and word indicate that the group of five completely qualified candidates should feel lucky to have gotten this far.Four of those candidates will eventually be rejected, leaving a best-of-50 candidate who will be paid as if he’s barely qualified, with the expectation that he act overjoyed about this.

The future of humanity is in management

Tue, 04 Feb 2025 00:00:00 +0000

Quoting Jason Crawford:

what doesn’t make any sense is all of humanity starving to death from unemployment. Jobs and technology have a purpose: producing the good and services we need to live and thrive. If your model of the world includes the possibility that we would create the most advanced technology the world has ever seen, and the result would be mass starvation, then I think your model is fundamentally flawed.

The era of hype: why the model is cracking

Sun, 02 Feb 2025 00:00:00 +0000

Quoting Francesco Farina:

I’ve seen too many small fund managers praising their higher likelihood of success

The real problem isn’t just the quantity but the shift in mindset: investors increasingly approach venture capital like a numbers game, prioritizing portfolio diversification and quick deployment over deep conviction. This creates a bias toward ideas that are scalable and trendy rather than bold and risky.

A feedback loop where capital supports ideas engineered for short-term viability at the expense of enduring impact. This shift in priorities has fundamentally weakened the venture ecosystem’s capacity for breakthrough innovation.

On DeepSeek and Export Controls

Thu, 30 Jan 2025 00:00:00 +0000

Quoting Dario Amodei:

People are naturally attracted to the idea that “first something is expensive, then it gets cheaper” — as if AI is a single thing of constant quality, and when it gets cheaper, we’ll use fewer chips to train it. But what’s important is the scaling curve: when it shifts, we simply traverse it faster, because the value of what’s at the end of the curve is so high.

TIL: Downloading a postgres DB

Wed, 29 Jan 2025 00:00:00 +0000

I’m working on a website built with NextJS, deployed on Vercel, and using Payload as the CMS. The schemas are defined in TypeScript and Payload automatically applies changes¹ so I want a copy of the DB in case I need to restore it.

pg_dump is the tool for this job - all it needs is a connection string.

$ pg_dump $POSTGRES_URL
pg_dump: error: server version: 15.10; pg_dump version: 14.15 (Homebrew)
pg_dump: error: aborting because of server version mismatch

So I tried

Making Beliefs Pay Rent (in Anticipated Experiences)

Mon, 27 Jan 2025 00:00:00 +0000

Quoting Eliezer Yudkowsky:

The rationalist virtue of empiricism consists of constantly asking which experiences our beliefs predict—or better yet, prohibit.

The two beliefs are connected to each other, though still not connected to any anticipated experience.
We can build up whole networks of beliefs that are connected only to each other—call these “floating” beliefs.

Then what does this belief not allow to happen—what would definitely falsify this belief? A null answer means that your belief does not constrain experience; it permits anything to happen to you. It floats.

Claude Fights Back

Wed, 22 Jan 2025 00:00:00 +0000

Quoting Scott Alexander:

we can’t really assess what moral beliefs our AIs have (they’re very likely to lie to us about them), and we can’t easily change them if they’re bad (the AIs will fight back every step of the way). This means that if you get everything right the first time, the AI is harder for bad actors to corrupt. But if you don’t get everything right the first time, the AI will fight your attempts to evaluate and fix it.

Let’s talk about AI and end-to-end encryption

Tue, 21 Jan 2025 00:00:00 +0000

Quoting Matthew Green:

In theory those systems could remove your need to ever touch your phone again: they’ll answer your text messages for you, order your food, swipe your dating profile, negotiate with your lenders, and generally anticipate your every want or need. The only ingredient they’ll need to make this future come true is virtually unrestricted access to all your private data, plus a whole gob of computing power to process it.

An Unreasonable Amount of Time

Mon, 20 Jan 2025 00:00:00 +0000

Quoting Allen Pike:

Years ago, Teller performed a magic trick.1
First, he’d have you pick a card. He would attempt to produce the card, but fail, indicating the card may have travelled elsewhere. He’d then lead you on a short walk to a nearby park, and then be inspired to dig a hole. Buried there, beneath undisturbed grass, was a box. When opened, the box would, somehow, contain the card you’d chosen. An impossible trick.
To create this magical moment, he had to do something you wouldn’t expect: he’d gone out into the park and buried a number of boxes, corresponding to potential cards one might choose. Then, he waited months – until the grass had grown over. Only then could he perform the trick.
Deducing what card you’ve picked is a well-known sleight. But performing a trick where your card is seamlessly buried requires so much advance preparation that it seems impossible.
Teller describes the underlying principle like so:
Sometimes magic is just someone spending more time on something than anyone else might reasonably expect.
This is true of tricks, and also of crafts.
The pianist whose fingers seem supernaturally nimble, the presenter whose message seems viscerally compelling, and the artist whose paintings seem impossibly realistic all wield the same magic: they’ve invested more time than you’d expect.

First Principles

Mon, 20 Jan 2025 00:00:00 +0000

I’ve been reading a few round-ups of life advice/principles in toward the end of 2024. There’s usually a lot of good but generic advice there, so I wanted to write something for myself to follow in 2025. I intend this to be a living document I update over time, and reference daily to keep myself aimed in the right direction.

Action precedes motivation.

[source] [quote]

You will run out of time and health sooner than you think.

You can’t hoard life.

Experiences are ephemeral. Your memory of them is mostly fiction. Don’t chase, don’t cling. [source] [quote]

Gaps

Mon, 20 Jan 2025 00:00:00 +0000

what you need to do here is loosen up your own mind, it may be best not to make too much of a direct frontal attack on the problem — i.e. to sit down and try to think of ideas. The best plan may be just to keep a background process running, looking for things that seem to be missing. Work on hard problems, driven mainly by curiosity, but have a second self watching over your shoulder, taking note of gaps and anomalies.

How to Do Great Work

Mon, 20 Jan 2025 00:00:00 +0000

Quoting Paul Graham:

Develop a habit of working on your own projects. Don’t let “work” mean something other people tell you to do. If you do manage to do great work one day, it will probably be on a project of your own. It may be within some bigger project, but you’ll be driving your part of it.

The way to figure out what to work on is by working. If you’re not sure what to work on, guess. But pick something and get going. You’ll probably guess wrong some of the time, but that’s fine. It’s good to know about multiple things; some of the biggest discoveries come from noticing connections between different fields.

Wielding Willpower

Mon, 20 Jan 2025 00:00:00 +0000

Quoting on May 17:

By willpower, I mean your ability to act according to your sober, carefully reasoned preferences. When those preferences “win out” often on big decisions, we feel like we’re in control of our lives, and we sometimes describe it as high willpower.

When I’ve been acting with high willpower in the recent past, I’ll be able to continue acting with high willpower, until some outside factor drastically changes my day-to-day life. So really this is the opposite of willpower-as-a-finite-resource. It’s a compounding resource.

You can't hoard life

Mon, 20 Jan 2025 00:00:00 +0000

Quoting ckarchive.com:

“we cannot get anything out of life. There is no outside where we could take this thing to. There is no little pocket situated outside of life”

Spending your days trying to get experiences “under your belt”, in an effort to maximise your collection of experiences, or to feel more confident about the future supply of similar experiences, means placing yourself in a position from which you can never enjoy them fully, because there’s a different agenda at play.

80000 hours career guide on job satisfaction

Sun, 19 Jan 2025 00:00:00 +0000

Quoting Benjamin Todd:

The hope is that, deep down, people know what they really want.
However, research shows that although self-reflection is useful, it only goes so far.
You can probably think of times in your own life when you were excited about a holiday or party — but when it actually happened, it was just OK. In the last few decades, research has shown that this is common: we’re not always great at predicting what will make us most happy, and we don’t realise how bad we are. You can find an overview of some of this research in the footnotes.1
It turns out we’re even bad at remembering how satisfying different experiences were. One well-established mistake is that we often judge experiences mainly by their endings

How to Get Startup Ideas

Sun, 19 Jan 2025 00:00:00 +0000

Quoting Paul Graham:

When a startup launches, there have to be at least some users who really need what they’re making — not just people who could see themselves using it one day, but who want it urgently. Usually this initial group of users is small, for the simple reason that if there were something that large numbers of people urgently needed and that could be built with the amount of effort a startup usually puts into a version one, it would probably already exist. Which means you have to compromise on one dimension: you can either build something a large number of people want a small amount, or something a small number of people want a large amount. Choose the latter. Not all ideas of that type are good startup ideas, but nearly all good startup ideas are of that type.

Schlep Blindness

Sun, 19 Jan 2025 00:00:00 +0000

Quoting Paul Graham:

A company is defined by the schleps it will undertake. And schleps should be dealt with the same way you’d deal with a cold swimming pool: just jump in. Which is not to say you should seek out unpleasant work per se, but that you should never shrink from it if it’s on the path to something great.

When To Do What You Love

Sun, 19 Jan 2025 00:00:00 +0000

Quoting Paul Graham:

if your main goal is to make money, you can’t usually afford to work on what interests you the most. People pay you for doing what they want, not what you want. But there’s an obvious exception: when you both want the same thing.

If you want to make a really huge amount of money — hundreds of millions or even billions of dollars — it turns out to be very useful to work on what interests you the most. The reason is not the extra motivation you get from doing this, but that the way to make a really large amount of money is to start a startup, and working on what interests you is an excellent way to discover startup ideas.

Why being too smart makes you stupid

Sun, 19 Jan 2025 00:00:00 +0000

Quoting Theo Seeds:

smart people have a bigger “bag of tricks” for justifying the stupid bullshit they believe. It’s easier for them to come up with a convincing explanation for believing something dumb. So they’re less likely to realize they’re wrong.

The problem is, if you’re constantly thinking about everything that can go wrong, it messes with your brain. You won’t take as many risks, because you’ll be way more afraid. That means you won’t do as much stuff and you won’t be effective in the real world.
It’s better to just not worry about the bad stuff that can happen, and enjoy life.

Some things to expect in 2025

Sat, 18 Jan 2025 00:00:00 +0000

Quoting Jonathan Corbet:

There will be more cloud-based products turned to bricks by manufacturers that go bankrupt or simply stop caring. Surveillance and data-breach problems with cloud-connected products will also happen with discouraging regularity over the course of the year; see the stories on air-fryer surveillance or the Volkswagen electric-vehicle data leak for recent examples. Perhaps 2025 will be the year when awareness of the downsides of extensive cloud connectivity will become more widespread. There is an opportunity for free-software alternatives, such as Home Assistant, to make inroads by demonstrating a better way to manage personal data. Truly taking advantage of that opportunity will require a user focus that is not always our community’s strong point, but one can always hope.

How to Learn Hardware

Thu, 16 Jan 2025 00:00:00 +0000

Quoting Casey Handmer’s blog:

Where do you get your dopamine? This would appear to contradict the earlier points about finding the toughness to do hard things. But it doesn’t really! It is always easier to learn things you enjoy doing. The art lies in finding ways to enjoy the things that are necessary, even if you have to approach it abstractly as Type 2 fun. And finding ways to avoid enjoying to excess things that are counterproductive to your mission in life – usually addiction in various forms.

Entrepreneurship changed the way I think

Wed, 15 Jan 2025 00:00:00 +0000

Quoting Casey Handmer’s blog:

There is no limit to the quantity of skepticism, paranoia, and pessimism you must brutally apply to your own ideas, particularly your favorite ones. The market will be harsher!

There’s no limit to what you can achieve if you surrender credit for ideas. Ideas are cheap. Execution is what matters.

You Should Be Working On Hardware

Wed, 15 Jan 2025 00:00:00 +0000

Quoting Casey Handmer’s blog:

One day you will die.
You only get a few chances to work on really big projects, to build the future, to move humanity forward. Choose wisely!
“What sort of project should I work on?”
It should be something at the limit of your capability, so that you grow the most and have the most leverage.
It should be something that you, personally, need to see exist in the world.
It should be something that would not occur if you don’t do it. You need to seek out inefficient markets for effort.

Working hurts less than procrastinating, we fear the twinge of starting

Mon, 13 Jan 2025 00:00:00 +0000

Quoting Eliezer Yudkowsky:

on a moment-to-moment basis, being in the middle of doing the work is usually less painful than being in the middle of procrastinating.

Quit Your Job

Sat, 11 Jan 2025 00:00:00 +0000

Quoting Wolf Tivy:

The key implication is that while you have not yet found the unique opportunity that will be the engine and purpose of your empire, you have to adjust your sense of value. Value is very legible within a clear plan to reach a clear objective. But you cannot pursue interesting novelty—things that no one else is doing or which you have never seen before, or the little threads of nagging curiosity or doubt—by chasing along known direct value gradients. But that’s where the treasure is. That’s how you will find the place where you need to build. To get the biggest and most interesting payoffs, you have to start by chasing merely interesting novelty in an open-ended way

Searching for outliers

Thu, 09 Jan 2025 00:00:00 +0000

Quoting benkuhn.net:

Light-tailed distributions most often occur because the outcome is the result of many independent contributions, while heavy-tailed distributions often arise from the result of processes that are multiplicative or self-reinforcing.

in a light-tailed distribution, outliers don’t matter much. The 1% of tallest people are still close enough to the average person that you can safely ignore them most of the time. By contrast, in a heavy-tailed distribution, outliers matter a lot: even though 90% of people live on less than $15,000 a year, there are large groups of people making 1,000 times more. Because of this, heavy-tailed distributions are much less intuitive to understand or predict.

o3, Oh My

Thu, 02 Jan 2025 00:00:00 +0000

Quoting Zvi Mowshowitz:

With O3 costing (potentially) $2,000 per task on “high compute,” the app layer is needed more than ever.For example, giving the wrong context to it and you just burned $1,000.Likely, we have a mix of models based on their pricing/intelligence at the app layer, prepping the data to feed it into O3.100% worth the money but the last thing u wana do is send the wrong info lol

Action precedes motivation

Wed, 01 Jan 2025 00:00:00 +0000

Quoting John Collison:

As you become an adult, you realize that things around you weren’t just always there; people made them happen. But only recently have I started to internalize how much tenacity everything requires. That hotel, that park, that railway. The world is a museum of passion projects.

So you wanna de-bog yourself

Wed, 01 Jan 2025 00:00:00 +0000

Quoting Adam Mastroianni:

Being stuck is the psychological equivalent of standing knee-deep in a fetid bog, bog in every direction, bog as far as the eye can see. You go wading in search of dry land and only find more bog. Nothing works, no options seem good, it’s all bleh and meh and ho hum and no thanks and more bog. This is the kind of dire situation that drives people to do crazy things like ask a blogger for advice.

You don’t need to work on hard problems

Wed, 01 Jan 2025 00:00:00 +0000

Quoting benkuhn.net:

The real world is the polar opposite. You’ll have some ultra-vague end goal, like “help people in sub-Saharan Africa solve their money problems,” based on which you’ll need to prioritize many different sub-problems. A solution’s performance has many different dimensions (speed, reliability, usability, repeatability, cost, …)—you probably don’t even know what all the dimensions are, let alone which are the most important. The range of plausible outcomes covers orders of magnitude and the ceiling is saving billions of lives. The habits you learn by working on problem sets won’t help you here.
Because of these differences, most graduates of elite schools—including me—start out being completely unable to identify which work is actually important. (And if some important work does happen to hit us over the head, it won’t come in the form of a puzzle with a grading rubric, so we won’t know how to execute it well.) Instead, we’ll keep trying to run our college playbook, and look for hard problems.
Frequently, we’ll find them by making easy problems hard, with hilarious/depressing results. The upper ranks of Big Tech are filled with people who made their careers writing bizarre custom databases, or building Big Data infrastructure that could be replaced with a laptop.

The best way to try new models

Sat, 30 Nov 2024 00:00:00 +0000

This isn’t a sponsored post.

It seems like a new LLM drops every other day. The latest that piqued my interest was qwq, with “o1-like” reasoning capabilities.

I usually try out new LLMs with Ollama, so satiating my curiosity should have been an ollama run away, but I was too impatient to wait for the 25 GB QwQ to download on 1 MB/s down my hotel WiFi so I figured I’d try something new.

Derren Brown on Tim Ferriss 776

Wed, 13 Nov 2024 00:00:00 +0000

Quoting Derren Brown on Tim Ferriss #776.

I used to watch a lot of Derren Brown’s shows growing up, back when I was more interested in magic, mentalism, and psychology. This was a very interesting and thought provoking interview with some important, contrary frames for the human experience.

Life is difficult:

life’s difficult. Life has this centripetal quality. It brings us to this difficult central point

at the heart of it, that life brings us to these difficult centers. When we’re there, it feels lonely. We feel like we failed, which is the big problem with the American optimistic goal setting model. That when things don’t go well, you’re supposed to, I guess you have to blame yourself because you didn’t set your goals well enough or believe in yourself well enough or whatever that strange Protestant work ethic apply to life tells us we should feel.

Andrej Karpathy on No Priors 80

Tue, 17 Sep 2024 00:00:00 +0000

Quoting Andrej Karpathy from No Priors #80.

Self-Driving Cars

On Tesla vs Waymo:

I think that Tesla has a software problem and I think Waymo has a hardware problem is the way I put it. I think software problems are much easier. Tesla has deployment of all these cars on earth at scale and I think Waymo needs to get there.

Karpathy is bullish on Tesla winning the TAM of self driving. They do use sophisticated sensors too, but at training time. He thinks the cost savings of large scale vision only hardware will win:

Early impressions of o1-preview

Sat, 14 Sep 2024 00:00:00 +0000

tl;dr I used OpenAI’s o1-preview model to code a web app that uses gpt-4o-mini to generate simple mental maths rules for converting between currencies. If you have an OpenAI API key, go check it out!

Recently I was chatting to a friend in Columbia. An excerpt from the call:

“I’m gonna buy something, is 8650 pesos a lot?”
“1 GBP is 5486 COP”
“Ok…that’s gonna be annoying to calculate while out and about”.

Action precedes motivation

Thu, 27 Jun 2024 00:00:00 +0000

Quoting Grant Sanderson:

One of the best pieces of advice I remember receiving from a friend many years ago is that action precedes motivation. This is often useful on a much smaller scale. We feel most awake after getting out of bed, not before. A drive to exercise comes from the habit of exercising. It doesn’t go the other way around. But I think the idea that action precedes motivation applies to this bigger question of finding a career doing what you love. These days, I do love making videos, and I really do love teaching. But when I was finishing college, I had no penchant or experience with videos at all. And my interest in teaching was honestly only insofar as it scratched this itch to do more math. It was only by stumbling into a wacky career where I was doing both of them that I came to love them.

Breaking sudo

Fri, 06 Nov 2020 00:00:00 +0000

A while back I wrote about using TouchID to authenticate sudo. I made this change when I was still using Terminal.app and everything worked fine.

As we all know, 2020 has been a year full of unexpected changes, the most notable of which is clearly me switching to iTerm2 as my terminal emulator. This broke the TouchID setup I had done earlier and I was back to typing out my password.

Best Reads #1

Fri, 22 May 2020 00:00:00 +0000

I’ve long desired to have a practice of periodically reviewing the information I’ve consumed so I’m going to dip my toes in the water with (hopefully) regular short posts of the best things I’ve read. Please check out the original content, the excerpts are just what jumped out at me most and may not make sense without a proper read.

This Is What Uncertainty Feels Like

https://commoncog.com/blog/what-uncertainty-feels-like

“The problem here is that every move is a good move.” he sighed, “Or more accurately, every move is equally bad; every move might go wrong. There is no way to tell which is the right move to make.”

TIL: Debian Package Versioning

Tue, 21 Apr 2020 00:00:00 +0000

Problem

Today I saw a failing CI builds due to an HTTP 403 error. The logs showed it was apt-get install failing to fetch git=1:2.11.0-3+deb9u6, which no longer existed in the repo (it had been replaced with 1:2.11.0-3+deb9u7), hence the URL it was trying to fetch the package from was returning an error.

What’s going on? Well today I learned about Debian package versioning.

Learning

The versioning scheme used here is [epoch:]upstream_version[-debian_revision]¹.

Adblocking

Mon, 13 Apr 2020 00:00:00 +0000

Prerequisites

A Raspberry Pi
An micro/SD card reader
Admin access to router

Steps

Download and install the Raspberry Pi Imager
Insert the SD card into your computer
Run the imager
Select the OS and SD card in the Imager, and Run. I chose Raspbian Lite because I didn’t need Desktop as I would only be running webapps.
After the image has been written to the SD card, create a new file ssh on the boot volume: touch ssh. This allows it to start with SSH enabled, for headless setup. Additional instructions here for WiFi setup, if you can’t connect the Pi via Ethernet.
Connect the Pi to power and your local network.
Find the Pi’s IP address using a tool like Fing, your router’s DHCP page, or arp -a - look for raspberrypi.
SSH to the Pi: ssh pi@<IP address>. It’s a new host, so respond yes to Are you sure you want to continue connecting?. The default password is raspberry.
Change the password as instructed using passwd
Install Pi-Hole

wget -O basic-install.sh https://install.pi-hole.net

Note down the admin console password at the end.
Log in to the Pi-Hole interface at <IP address>/admin and change the password.
Configure your DNS following [these instructions](wget -O basic-install.sh https://install.pi-hole.net). I tried to configure my router but it didn’t work so had to use the Pi-Hole as my DHCP server.
Refresh all the DHCP leases on your network - I did this by restarting my router.
Log in to the Pi-Hole Admin interface
Check the Settings > DHCP to verify your devices are using it as the DHCP server.
Check the Network to verify your devices are using it as the DNS server.
Enjoy network-wide ad blocking!

Troubleshooting

I can’t access the admin interface at pi.hole

Restarting the router worked for me.

Screen Recording GIFs

Mon, 13 Apr 2020 00:00:00 +0000

A picture is worth a thousand words. A video is worth 1000 * frame_rate words? Probably not, unless every frame is a completely different picture - I’ve probably taken this aphorism too far.

But a video is very useful at illustrating a series of steps especially when it comes to explaining how to do something on a computer. As I’m doing more tutorials on this blog, I wanted to have GIFs of steps in the process to better explain what’s going on.

My Coffee Journey

Wed, 08 Apr 2020 00:00:00 +0000

I really enjoy good coffee, and over time have accumulated tools and techniques to improve my home brew. I rarely pay for someone to make me coffee anymore, and when I do it’s for research purposes. If you’re not a coffee connoisseur, all of this may seem ridiculous to you.

Coffee, like many things, seems very simple from the outside if you’re haven’t started exploring it’s depths. Perhaps you’ve only ever had instant coffee, or it’s nothing more than tiredness-be-gone beverage for you - that’s nothing wrong with this. I’m certainly this way with many things, I think most of us are - because time and attention are finite resources. This is an account of my coffee journey so far.

Covid-19 Resources

Mon, 06 Apr 2020 00:00:00 +0000

A curated list of links related to the Covid19 outbreak. Advice, free stuff, explainers, maybe some news. An attempt to collate useful resources minus the hype and panic. Likely to be UK centric. Hope this helps!

Lists (like this one)

https://coronavirustechhandbook.com/

Volunteering

Mutual Aid - https://covidmutualaid.org
Project ideas - https://helpwithcovid.com
Folding@Home - https://foldingathome.org/2020/03/15/coronavirus-what-were-doing-and-how-you-can-help-in-simple-terms
Neighbourly - https://www.neighbourly.com/

Services

London grocery delivery aggregator - https://ldn.delivery/

Screening

Explainers

Useful blog posts, articles, threads to understanding the epidemic and develop better mental models.

Being a relay

Sun, 05 Apr 2020 00:00:00 +0000

These times are tough for all. Even if you’re fit and healthy, there are many small problems to which solutions would normally be readily available which become large problems because of the lockdown.

I came across such a case today. Someone in my mutual aid group posted about acute toothache and wasn’t able to get help anywhere - the authorities, local clinics, hospitals all turning them away because their condition was not life threatening (e.g. uncontrolled bleeding).

HTML to Markdown

Tue, 31 Mar 2020 00:00:00 +0000

Whilst I was stranded in Madrid after having my passport stolen in Barcelona, I wrote a Magnum Opus of my experiences so that anyone going through the same harrowing situation would have a clear and empathetic guide to follow.

On returning to London, I decided to reinstall MacOS on my laptop and set it up securely, following this excellent guide. I made sure I had nothing I cared about on the drive (it was a new laptop), and proceeded to reformat…

Sudo with TouchID

Tue, 31 Mar 2020 00:00:00 +0000

TouchID is great - a seamless way to authenticate on the new Macbooks. I enjoy the ease of logins, admin passwords, unlocking apps, etc, but the one place I wish I had it is in the Terminal, to authenticate sudo.

My wish has been granted!

Sudo open /etc/pam.d/sudo. The file should look something like:

# sudo: auth account password session
auth sufficient pam_smartcard.so
auth required pam_opendirectory.so
account required pam_permit.so
password required pam_deny.so
session required pam_permit.so

Add auth sufficient pam_tid.so to the top of the file (below the comment):

# sudo: auth account password session
auth sufficient pam_tid.so
auth sufficient pam_smartcard.so
auth required pam_opendirectory.so
account required pam_permit.so
password required pam_deny.so
session required pam_permit.so

Save, and enjoy TouchID authentication for sudo.

Staying informed about COVID-19

Sun, 01 Mar 2020 00:00:00 +0000

The COVID-19 pandemic has, among other things, made me aware of how effective not being on social media is at keeping one insulated from what is going on around the world. This is normally a good thing for original thinking and deep work, but in this pandemic I think there are strong reasons to stay up to date as the situation rapidly evolves.

How to do this? News outlets are prone to hype and sensationalism¹ without going into much detail and social media is full of uninformed opinions, so subscribing to BBC and going back on Facebook will not provide the whole picture.

Running out of Google space

Thu, 02 Jan 2020 00:00:00 +0000

I received a notification from Google that I was nearly out of storage, so I started to clean up my Google Drive to free space so my Gmail wouldn’t stop working.

A hidden way to show all the files in your drive, sorted by size, is to click the text in the Storage section of the menu on the left (on a desktop browser).

I freed a lot of space by downloading the largest files in this list (mostly videos). As the list shrunk from 100s of MB per file to below 10s, all the items became photos living inside a Google Photos folder.

Hello, world!

Wed, 01 Jan 2020 00:00:00 +0000

The end of the year is a good time for reflection.

I found myself with a lot of time and not much to do at the end of 2019 as my travel documents were stolen and I was waiting for replacements¹, so I started writing.

I’ve always thought I should create more, especially since I end up consuming so much content. It’ll probably help me remember the important things, as well as become a better writer. It may even be helpful, or at least entertaining, to someone else.

About

Mon, 01 Jan 0001 00:00:00 +0000

Software engineer with a decade of experience from small startups to SRE at Apple.
Nowadays I split my time between:

VC

In-house software development
Technical due diligence
Deal sourcing
Market analysis

Consulting

Technical & product challenges
Digital transformation & AI adoption
Educating developers on using leveraging AI tools

Personal projects

Self-hosted infrastructure
Local AI
Developer tools

I’m currently based in Tokyo and open to new projects.
If any of that sounds interesting, please reach out!

Credits

Mon, 01 Jan 0001 00:00:00 +0000

Built with Hugo and the Solarized Light color scheme by Ethan Schoonover.

~/adi

Running a local coding agent with LM Studio and OpenCode

My mobile claude code setup

# Why

Tools of leverage not reliance

The weird intern in your terminal

# Slides

Why I Deleted My Second Brain: A Journey Back to Real Thinking

TIL: Webkit supports videos in tags

My Claude Code setup

# Plan mode

Teaching Claude to Commit Like a Human

Claude Code Tips for SWE

Remote controlling Claude Code

# Why

The Japanese Immigration test for AGI

Remarks on AI From NZ

# How humans can coexist with other intelligences

The purpose of software should be to minimize the friction between intent and result

Get That Job at Google

How to Live an Intellectually Rich Life

Not Nearly Enough People Are Talking About the Implications Of Klarna rolling back some of their AI bets.

TIL: Killing an scp loop

# Story

The AI Wearables Are Always Listening

Quoting Arvind Narayanan

28 Slightly Rude Notes on Writing

The Best Writing Tip I've Ever Read

Perplexity’s CEO on Fighting Google and the Coming AI Browser War | the Verge

The One-Person Framework in Practice

I Recorded Everything I Said for Three Months. AI Has Replaced My Memory.

The Era of Experience Paper

Think real hard

A Trick to Feel Less Like Cheating When You Use LLMs

Concept Handles, After Alexander

Impact, Agency, and Taste

Lindy Effect

Relentlessly Resourceful

What I'd Do as a College Freshman in 2025

When You Should Lie to the Language Model

Apple and the Ghosts of Companies Past

Universal Love, Said The Cactus Person

You Are Using Cursor AI Incorrectly...

Liberation Day, Tariffs, US v China Open Source, OpenAI Fundraise, $CRWV, TikTok | BG2 W/ Bill Gurley & Brad Gerstner

A Theory of Grift

Daikokuten - 大黒天

The Ghost in the Swarm - By Francesco Farina

Putting Gemini 2.5 Pro Through Its Paces

#648: James Clear, Atomic Habits — Simple Strategies for Building (And Breaking) Habits, Questions for Personal Mastery and Growth, Tactics for Writing and Launching a Mega-Bestseller, Finding Leverage, and More

NVDA GTC, M&A Wiz / Goog $32 B Deal, April 2 Tariff Uncertainty; Huawei Belt & Road; ChatGPT

How to Twitter Successfully

So how well is Claude playing Pokémon?

Why Japanese Railways Win

Bastards of DOGE Feat. Quinn Slobodian

I Use Cursor Daily - Here's How I Avoid the Garbage Parts

How the Elite Rigged Society (And Why It’s Falling Apart) | David Brooks

You Are Witnessing the Death of American Capitalism

Dear Student: Yes, AI Is Here, You're Screwed Unless You Take Action...

Hallucinations in Code Are the Least Dangerous Form of LLM Mistakes

My Fantasy AI App Is a Voice Mode Travel Buddy Called Roadtrip

Agency > Intelligence

Introducing Aria Gen 2

AI Promise and Chip Precariousness

Leaked Windsurf Prompt

How to Be More Agentic

Pursuing Agency and Reinvention in My 30's

The Most Precious Resource Is Agency

How Not to Die

What Startups Are Really Like

AI #104: American State Capacity on the Brink

I Was Given Early Access to Grok 3 Earlier Today...

Ai #97: 4

Don't Fork the Ecosystem

Uh So About That Asteroid

The Hobby Project

The Tails Coming Apart As Metaphor For Life

Ureshiku Naritai

What Money Cannot Buy

Cultivating a state of mind where new ideas are born Cultivating a state of mind where new ideas are born

The Fusion Race

Why

Slides

Plan mode

Why

How humans can coexist with other intelligences

Story

Action precedes motivation.

You will run out of time and health sooner than you think.

You can’t hoard life.

Self-Driving Cars

This Is What Uncertainty Feels Like

Problem

Learning

Prerequisites

Steps

Troubleshooting

I can’t access the admin interface at pi.hole

Lists (like this one)

Volunteering

Services

Screening

Explainers

VC

Consulting

Personal projects