Brad Williams

Contra the peddlers of Progress and workaholism, new technology has DIMINISHING returns in the only thing that matters: inner peace. Data point: never has the West had such advanced tech (and large GDP), and never has it been generating more widespread spiritual misery. And by the materialists' own shallow standard of improving the physical world, technocapitalism is speeding humanity toward failures that are the stuff of apocalypse: the death of the ocean, world war (now with drones!), economic meltdown.... Their ultimate goal at the end of this great exercise in Progress? Leave Earth, and replace humans with computers running "consciousness" software.

Suffering is an appearance *inside* of peace and bliss, the eternal state.

When 'here' becomes 'Everywhere' and 'Now' becomes 'Always', then One has succeeded. Wu Hsin

There's nowhere to reach because wherever you arrive - it's always here and now. So why is everyone in a hurry as if there's a better place to get at? Habit and mind energy. Be aware of thoughts and realise you are present. Here and now nothing is missing. You are in heaven.

Alignment, we have a problem.

@AdamTornhill Feels much easier because I'm not burning concentration on all the details.

After 30+ years as a programmer, all my code is now written by agents. And I love it. But there’s a flip side: Software development is definitely more mentally effortful with agents. Anyone else feeling that?

@ThomasODuffy @aaronjmars I had 4.6 do two oopsies like that on me, so I'm guessing nothing's new with 4.7

@aaronjmars Be careful in how you use Claude Code Opus 4.7. Backup everything regularly. This happened today and was unrelated to my request.

opus 4.7 is genuinely the best & most intense model i've ever tried started using it & felt disappointed like everyone. then you read the recommendation & how it should behave : 'The model will not silently generalize an instruction from one item to another, and will not infer requests you didn't make.' it needs more information from you, more context, but when you feed the right problem & give him directions, he will fucking ship. which makes me really scared for Mythos, bc i already feel like 4.7 is a 2x improvement over 4.6 (on the right conditions)

@jesse_vermeulen Switch laptops

honest question: what do people do during the 5-10 min while Claude is running?

What you need will come to you, if you do not ask for what you do not need. Yet only few people reach this state of complete dispassion and detachment. It is a very high state, the very threshold of liberation.

@RKelanic The regime has already lost, dead-enders are just being dead-enders. "Enduring", is that a new euphemism for decomposing?

Blockading Iran will not compel its surrender. After enduring 6 weeks of war, Iran’s leaders won’t suddenly decide, “Oh, we can’t use our ports now. We better meet U.S. demands.” Trump can’t escalate the U.S. out of this mess. The war failed — now it’s time to cut losses. @defpriorities

@SchopenhauerNow Upanishads

Name the book that rewired your brain.

@DrNeilStone The regime is militarily defeated, but not politically because it's a bunch of psychopath dead-enders. We see it all across the ME: Islamicists aren't rational enough to surrender, they'd rather lose everything they have AND everything their people have.

It took 6 years to defeat Nazi Germany Who seriously believed it would take just 6 weeks to defeat the Iranian regime

@theoliverxp @Tradermayne Fourth Silo book? How does it stack up to the first three?

@Tradermayne If you haven’t yet, real Silo Series - don’t skip the 4th book… I’m a fan of Red Rising although I’ve been saving the final book forever as I don’t want it to end Silo is on the same level Don’t watch the series though, not the same

I just finished the Red Rising trilogy and, omg, one of the best fiction stories I've ever read. I flew through these three books in under a month. I only wish there was more. What are some other series that you consider on the same level?

@charliermarsh I keep seeing this take, but I'm still just using CC to perform the reviews and I almost never look at code anymore. That said, I drive CC pretty hard to review the same PR several times and to reconsider the design etc.

Tragically I am continuing to find that the most effective guardrail against slop is extremely talented engineers doing very thoughtful, human code review

@PositivFuturist It's still not great at understanding speech. That will be a huge unlock.

Still can’t believe computers can talk now.. Most unbelievable sci-fi development ever and most of the discourse is people poking holes in it.

@Dan_Jeffries1 I'm usually using one, and absolutely babysitting it. No idea how not-babysitting could possibly be a good idea for real work.

I honestly don't understand how anyone runs agents in parallel on serious codework without babysitting every one of them. Caught GPT 5.4 about to rip Bun out of my entire monorepo because it hit an Ink compatibility issue during a smoke test. The agent's "fix" was to swap the runtime for the entire monorepo. If I hadn't been watching that terminal, it would have. These little magic machines we call LLMs are wonderful but when I read these crazy ass policy statements like Superintelligence is so close and we aren't ready and lets change the whole social contract of countries and let's tax robot labor and go full UBI with zero fucking evidence that anything is actually happening here in real life that requires this kind of societal level surgery, I don't know what the hell people are smoking because these things still make cascading stupid decisions that compound every single day.

@mohansson @HodlMagoo Regime's been getting dusted for 34 days, but now they're about to do "game-over" on Americans. 🤡

@HodlMagoo Any idea what altitude these F15/F16/F18s were flying at? If Iranian AD/SAMs can target B2/B52s, it’ll be game over for air superiority..

The only question people should be asking today is why all of a sudden are the Iranians able to down U.S aircraft

@a_protsyuk @ohmypy In this timeline, code needs to be human-readable less every day.

@ohmypy the 512k lines is the actual story. we're building critical dev infrastructure that's too large for any human to audit. at some point the question isn't 'did it leak' but 'can anyone actually understand what they're running.' tbh that's true of most AI tools in prod.

I couldn't care less about Claude Code's source being leaked on npm. What terrifies me is that it's 512,000 lines of TypeScript code. HALF A MILLION lines of code for what's essentially a glorified API wrapper. I think the crucial point in our reality when we took the wrong turn was the invention of JavaScript. And we cemented our path to doom with the invention of TypeScript. Half a million lines of code. Dear Lord, have mercy on us.

@cramforce AI will maintain it.

To quote from my keynote at Vercel's internal offsite: Software is free as in puppies. It will pee in your bedroom and eat your furniture. The weight of every line of code is real. We will need to maintain it. We will need to port it. It goes into the context window. And somebody in this room will get paged at 2am because it did something unexpected

@JaredOfAI @garrytan No human can write good what-we-used-to-call specs for directing 37k new lines of code per day. Clearly the AI is filling tons of details.

the 23 single-commit fire-and-forget sessions out of 32 total is the number people should pay attention to. that's not pair programming, that's dispatching whole different skill set. you stop optimizing how to write code and start optimizing how to spec tasks clearly enough that the agent gets it right first try

Absolutely insane week for agentic engineering 37K LOC per day across 5 projects Still speeding up

@PDoomOrder1 @RizwanB38953183 @elonmusk @Rothmus Greed for power guarantees this race will proceed full speed. This is how a planet of egos operates. You cannot talk a "me/mine", or 7 billion, out of pursuing "more."

@RizwanB38953183 @elonmusk @Rothmus Don't develop artificial super intelligence. Right now no one knows how to do alignment and alignment is necessary to avoid catastrophe. The labs admit that they do not know how to align superintelligence yet race towards it anyway.