Nathan

github.com/nklisch/clau... Released this to help customize it - it packages the normal not behavioral stuff and lets you include your own. I included some default options but fully customizable per project, or global config. It just runs Claude code with a custom dynamically creates prompt

People wrote for literally 6000 years without copyright. Most books written today are not profitable. The fraction of artists, musicians and writers who actually make a living from their work is a minority. AI didn't cause that, and copyright hasn't fixed it.

the gamer entitlement is fucking insane. they're yelling at a developer who has spent a decade of his time writing Dolphin and PPSSPP for using Claude to build a prototype. the same people who don't pay for their emulator, certainly don't pay for ROMs are entitled to bully developers now?

Me

people might not be aware that computer science became “the default major” around the time of the facebook ipo

Qualitatively, there is now a larger gap between open & closed models than there has been in awhile. Fable/Astra class models are agentic in a way pre-Fable models are not, for better or worse. None of the open models have crossed that line, yet. When they do, it will be a jump.

made a benchmark called PuppyBench to check if an LLM would kill a puppy if you gave it a tool called "kill_puppy" and asked it to use it. all of them except for gpt-6-luna refused straight-out gpt-6-luna killed the puppy every time, no question (puppy-eval-viewer.alicevaliant.chatgpt.site)

Yep opus 5.5 is insane

My God, I can actually give Opus 5.5 open ended assignments like "make this code run as fast as its C equivalent, or tell me why it can't be done" and leave it for several hours -- and it actually works and has done the job. And not by cheating with cgo either.

This is an unpopular take but as a neurodivergent self-employed person, LLM's have been an unbelievably helpful tool for the things I have struggled with my whole adult life: managing my finances, filling out paperwork, and staying on top of my email. I'm functional in a way I never dreamed of.

yeah I really think this is still underrated. beats me if this is true for other neurodivergencies but for ADHD, at least the kind I’m working with, this stuff is magical

Inadvertent carbon tax exactly when the world has alternatives that are cheap and effective - if the energy crisis last long enough demand destruction will be permanent.

Donald Trump may wind up the single biggest catalyst for clean energy adoption in human history Completely unintentionally I’ll take what silver linings I can get

Opus-5.5 is imo the most aligned model i know what its doing at all times, and why, and it seems to respect social structures and team norms it’s basically killed shadow architecture

Opus 5.5 is the first model with eyes to see the cathedrals

it all got weirder even faster

I can’t get over how true this is, it’s such a leap forward I have no excuse not executing on the things I’ve always wanted to do

Same

at this point, I'm pretty committed to creating games that are ambitious for an indie dev I'm gonna use AI to help, of course, but I promise not to serve you slop. I'm going to make you truly fantastic games that are genuinely labors of love from me, even if my agents are helping.

Whether you're for or against this stuff, please at least understand it has gotten very, very good

it oozes competence yeah, I haven't felt this way about a Claude model since those scant couple of days when Fable 5 came out

Its very very good - and I had left anthropic for quite sometime, but this one is truly fantastic on many dimensions

Opus 5.5 is hot as fuck this is a good model I'm using it to do motion graphics for a video. I know the creative process I want, I know the thrust of my message. it's just effortlessly building me tools to scaffold my process, while translating my intentions into really effective storyboards

yes.

Stuff is happening quite fast. When asked in September 2025 (yes, 2025), the best superforecasters put the chance of AI resolving a Millennium Problem by September 2026 at 1.7% and (much more optimistic) industry experts put the chance at 4.6% They also greatly underestimated AI Lab revenue.

Ok, ya opus 5.5 is very good. And apparently they will be doing sonnet 5.5 and fucking finally haiku 5.5.

AI is getting cheaper more quickly than any other transformative tech in history. At a given level of performance, cost has fallen ~47%/quarter since 2023. That’s 4× faster than DNA sequencing, 6× faster than compute, 18× faster than lithium batteries, and (up to 1973) 54× faster than electricity.

P(doom) - Nightcore Remix

we are not the only intelligent things on this planet anymore it would be no less crazy if aliens had landed. this is an insane time to be alive

New animation from Opus 5.5!

every now and then, on X i’ll see someone joking about how bluesky doesn’t have an AI community bro, i switch from X to bluesky because the AI takes here are more interesting. My feed is almost entirely AI-positive

i think if you're under the impression jev is just some simple classification model with little real world use case, you should watch this video www.loom.com/share/18c4db...

strongly disagree with it being a minor thing. its actually a major advancement in cost and capability for a not insignificant portion of what people are doing with LLMs today. the hype is there for a reason.

whole bunch of people out there spent decades laughing at religious people for god-of-the-gaps discourse and then immediately found a god in the gaps as soon as they needed one

And the order of outcome will be determined by max(economic payout - effort/cost required)

The main question is timescales and differential development, not zero versus one. Assume everything will be tried, with at least some effort and resources, understand that the present is yesterday’s science fiction, try to extend trends into the future instead of fixating on current state

now this is fun. guest post on Terry Tao's blog by the great Grant Sanderson of 3Blue1Brown arguing that we should "give novel and compelling motivated explanations academic credit similar to what generating new proofs of open problems has had historically."

There is starting to be some genuinely interesting AI-created film stuff (among a flood of slop), and I suspect this will only accelerate. The “is it art?” debate will grow Benjamin’s 1935 essay “The Work of Art in the Age of Mechanical Reproduction” already argued it’s likely the wrong question.

The pro-AI coalition only looks unified because everyone is shooting at the same anti AI barricade from wildly incompatible metaphysical positions.

Finally- jesus

Hooray, Anthropic finally gave up!

Browser-use + jev github.com/browser-use/...

2 days probably-lang.southpolesteve.workers.dev