following my curiosities | prev @chapterone @decentralisedco @biconomy
Joined August 2015
- Tweets1.8K
- Following1.4K
- Followers1.7K
- Likes6.8K
nishil retweeted
go fuck yourself @sama
claiming that you solved navier stokes because 10000 agents ran in circles for 88hours on a multimillion dollar gpu cluster to formalize in lean a blowup case under controlled external forcing is pure scientific vulgarity the clay mathematics institute millennium prize does not ask m whether you can artificially force a singularity in a fluid by injecting an ad hoc smooth external forcing term f(x,t) to twist the vortex until it breaks the real problem questions the fundamental stability and global smooth existence for 3dimensional incompressible euler & navier stokes equations under natural conservation laws and viscous dissipation alone using a mathematical loophole on forced equations to parade a century old victory is a major conceptual scam Altman
technically & epistemologically what you present as an agi breakthrough is nothing more than bruteforce combinatorial autoformalization the ai did not understand fluid mechanics it simply navigated a continuous search space previously mapped out and constrained by the monumental work of human mathematicians like tristan buckmaster/ levent alpöge / diego córdoba or tarek elgindi coordinating 10000 agents to check the logical consistency of a 100 page proof via lean is a software engineering feat and computational parallelization triumph not an intrinsic scientific discovery it is the victory of the compute bulldozer over abstract human intuition repackaged for the public as a higher mathematical consciousness
to this theoretical imposture you add a disgusting ethical and industrial cynicism taking advantage of private codex sessions and informal preprints from academic researchers to siphon their research leads and then trying to redact or erase the contribution of levent alpöge under the pretext that he works at rival anthropic is intellectual serfdom openai behaves like a feudal lord of silicon appropriating the cognitive subsistence of independent scholars threatening their careers behind closed doors if they protest and turning community academic labor into a privatized pressrelease
this entire staged event serves a desperate financial agenda in a pre ipo panic facing the slowdown of scaling laws and growing investor skepticism over the profitability of foundational models openai needs to manufacture an artificial sputnik moment claiming to solve a millennium prize without immediately submitting the proof to traditional peer review means using the prestige of fundamental mathematics as cheap marketing fuel to inflate a delusional valuation!!!
real science is not a clout chase on social media or a compute spike spent to rob the clay mathematics institute it is a quest for elegance physical truth and universal rigor to decode reality true artificial intelligence will not emerge from hostile corporate takeover of academic work hidden behind computational bruteforce but from architectures capable of generating new conceptual paradigms by masquerading constrained formalization as the collapse of physics greatest mysteries you did not solve navier stokes you only proved how far silicon valley will go to prostitute scientific integrity for capitalist spectacle
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.
The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.
The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
nishil retweeted
If you know how to live you have nowhere to live in India
'I can now live where I want, in clean air, running through nature, walking through urban beauty, living in private austerity and public splendour. This means I have to leave India. Because most of the country is a mofussil hellhole in the smog.'
Read free on my Substack.
manujoseph.substack.com/p/if…
brewing one for prediction markets next 👀
Was skimming through my old articles when I found this:
An old article where I called out the launch of an open standard for stablecoins, 6 months before OUSD was launched.
Had written about the why in my substack article titled: THE VISA EFFECT 💳
Go read it!
open.substack.com/pub/thetok…
Was skimming through my old articles when I found this:
An old article where I called out the launch of an open standard for stablecoins, 6 months before OUSD was launched.
Had written about the why in my substack article titled: THE VISA EFFECT 💳
Go read it!
open.substack.com/pub/thetok…
nishil retweeted
For the past 3 years my elite friends that are either working in AI or living in San Francisco have been telling me adamantly we are 1-2 years from AGI/ASI singularity and absolutely everything is about to change.
Meanwhile my intuition has been that no we’re not close to anything like humans not needed for intelligence. Like nowhere near I dont even see any whites on the eyes of this digital high IQ terminator, this dude is far on the horizon with the little mountains over there.
Stark division of opinions throughout. My intuition Ive learned to train and listen to it, its my most powerful weapon. But I am a humble weaponholder, I recognize the achievements and abilities of these friends, in some cases of the most elite of them, far outweighs what I have achieved. If my opinion clashes so much with people of this caliber I also sit my ass down and listen.
Its been enough time now that we’ve gone through their initial timelines. And somehow it feels like we have both been right?
Clearly something revolutionary is going on, the digital/software world is being terraformed in front of our eyes and taking huge swaths of the economy with it.
But I have been, so far, clearly correct. These things are still closer to a tool rather than replacing the tool holder. I still dont see it on the horizon. I dont know if its what Ilya S says that the data is gonna be the constraint. That feels right to me. There are only so many non linear intelligence vectors to train upon. There are still humans finding new ideas and innovating above and beyond what is available to train on online. I like to believe I am one of them and I dont feel like I am anywhere near saturation or that it will happen anytime in my lifetime.
Yan LC has his belief that LLMs are capped which was well articulated, now though combined with world models I don’t know if thats the constraint.
Data might be the boundary that will saturate first and the innovation needed for the next stage might be creating useful artificial data, much like for a long time financial model training has involved creating fake training data that has the attributes needed to program alpha.
Ultimately this is all a pursuit for more alpha. Its a bit unnerving that even though I have been correct that the timelines on us being replaceable meat boxes arent correct, its impossible for me disprove the notion that its at the next corner we are about to reach. I feel like we will just keep going one block down the street at a time with baited breath because clearly something is getting closer though holding your breath for how long its actually gonna take is gonna asfixiate us.
So lets take a breath. Theres a lot of trauma under the surface from the rate of change accelerating. Every single one of us feels it. The fascia are being pulled and we are emotionally masking it under our skin. We know something is under way. But those of us with strong awareness and with the lights switched on aren’t going to stop being needed anytime soon- I don’t care how many Elons, Darios, and Leopolds say otherwise.
nishil retweeted
Today we’re launching @decawork backed by @ycombinator : one place for IT to deploy, govern, and maintain the internal AI agents running across a company.
Every team is building AI tools/agents now. They go live with real access, no standard way to ship, nobody watching.
With Decawork, every agent - built in Claude Code, Codex, or anywhere else - ships with its own identity and scoped credentials. IT signs off before it's live, sensitive actions wait for approval, and every run is watched from first to last.
I built AI at NVIDIA, deployed at OpenAI and Meta. Aman (@arxneo14) built Barclays' AI compliance platform used by internal teams, then shipped agents to 100K+ monthly users. We lived this and quit to solve it.
Internal AI agents/tools piling up at your company? Chat with us founders[at]decawork[dot]ai or decawork.ai
reading about hayek, bastiat and how economies are most efficient in a bottom up free market, as opposed to when decisions and control comes top down
it’s making me question how much of governance and control is a necessity and what part of it is just hindrance
nishil retweeted
Second edition this Sunday, Aug 2nd 5-7pm @BLRxZo
Same buzzers, same negative points, an entirely different board.
Luma link in the thread - please register
Been into quizzing since I was a kid..college is where I actually found the tribe: quiz clubs, people who taught me how questions get built, not just answered
Then life happened and it faded - the way things do; the things that aren't quite hobbies; the things that aren't passions you'd die for; the things that just sit somewhere in between
Trying to rekindle that and what better place to host this than at our @BLRxZo
This Sunday, 5th July - 5pm at Zo House, Koramangala - Luma linked in the tweet below
P.S. don't worry - quizzing isn't like trivia night, where it's either you know it or you don't. That's just a badly written quiz. It's about connecting dots between random pieces of knowledge already sitting in your head
welll guysss shut up with your gatekeeping this fucking mid everyone would know cafe and your cringeworthy worshipping of a fucking AI model improvement and the cocksucking of your love for latent, there’s a crisis going on in the country, let’s fucking give fucks about that for a bit
nishil retweeted
Anthropic and OpenAI are like TikTok for enterprise data, except the feed is your work and the engagement signal is your judgment
How is there no consumer facing product yet that makes it easy to store and access your thoughts and research?
Seems like a no brainer for startups to chase this problem. A product with UX done right is an immediate sell to strategists, creators, founders, researchers.
What am I missing?
LLM Knowledge Bases
Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:
Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.
IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).
Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.
Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.
Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.
Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.
Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.
TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.
just watched common side effects, such an interesting show
really tempted to get a tortoise and name it socrates 🐢
