nicole grossmann

ml @ georgia tech | math @ columbia | founding team @ Enter AI

[nicole grossmann @nicgrossmann]

@nickscamara_ can't miss it!

[nicole grossmann @nicgrossmann]

🔥☄️

[Nicolas Camara @nickscamara_]

Vercel 🤝 Firecrawl

[nicole grossmann @nicgrossmann]

Try it now here: https://t.co/tv3T20sFR3

[nicole grossmann @nicgrossmann]

I built an open source caffeine decay curve simulator (to secretly maximize how many cups of coffee I can have each morning) ☕️

Features:
- # of coffee cups; caffeine per cup
- metabolism speed selection
- anxiety/jittery effects
- toggle for sugar

100% Open Source :)

[Tri Dao @tri_dao]

After some mathematical rewrite, turns out all of transformer is a series of gemm + epilogue. Given a few optimized primitives, LLMs (and novice humans) can write speed-of-light kernels for all transformer ops!

[nicole grossmann @nicgrossmann]

i love math!!!!!!!!

[OpenAI @OpenAI]

Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946.

For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids.

An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.

This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

[nicole grossmann @nicgrossmann]

i want candy!!! oops i meant macminis 🤓🤓🤓🤓

[nicole grossmann @nicgrossmann]

apparently the importance of being earnest is hot again! special thanks @oscarwilde

[Y Combinator @ycombinator]

“When founders are both formidable and earnest, they're as close to unstoppable as you get.”

Paul Graham on earnestness: https://t.co/gxNeNZMuew

[nicole grossmann @nicgrossmann]

Peter Thiel in Brazil was epic! 🇧🇷🇧🇷🇧🇷 So special to see a brazilian company host one of the biggest voices of entrepreneurship and innovation (at home)!

[nicole grossmann @nicgrossmann]

@nickscamara_ cool

[nicole grossmann @nicgrossmann]

Poof 💨

[Elias Al @iam_elias1]

MIT just made every AI company's billion dollar bet look embarrassing.

They solved AI memory. Not by building a bigger brain. By teaching it how to read.

The paper dropped on December 31, 2025. Three MIT CSAIL researchers. One idea so obvious it hurts. And a result that makes five years of context window arms racing look like the wrong war entirely.

Here is the problem nobody solved.

Every AI model on the planet has a hard ceiling. A context window. The maximum amount of text it can hold in working memory at once. Cross that line and something ugly happens — something researchers have a clinical name for.

Context rot.

The more you pack into an AI's context, the worse it performs on everything already inside it. Facts blur. Information buried in the middle vanishes. The model does not become more capable as you feed it more. It becomes more confused. You give it your entire codebase and it forgets what it read three files ago. You hand it a 500-page legal document and it loses the clause from page 12 by the time it reaches page 400.

So the industry built a workaround. RAG. Retrieval Augmented Generation. Chop the document into chunks. Store them in a database. Retrieve the relevant ones when needed.

It was always a compromise dressed up as a solution.

The retriever guesses which chunks matter before the AI has read anything. If it guesses wrong — and it does, constantly — the AI never sees the information it needed. The act of chunking destroys every relationship between distant paragraphs. The full picture gets shredded into fragments that the AI then tries to reassemble blindfolded.

Two bad options. One broken industry. Three MIT researchers and a deadline of December 31st.

Here is what they built.

Stop putting the document in the AI's memory at all.

That is the entire idea. That is the breakthrough. Store the document as a Python variable outside the AI's context window entirely. Tell the AI the variable exists and how big it is. Then get out of the way.

When you ask a question, the AI does not try to remember anything. It behaves like a human expert dropped into a library with a computer. It writes code. It searches the document with regular expressions. It slices to the exact section it needs. It scans the structure. It navigates. It finds precisely what is relevant and pulls only that into its active window.

Then it does something that makes this recursive.

When the AI finds relevant material, it spawns smaller sub-AI instances to read and analyze those sections in parallel. Each one focused. Each one fast. Each one reporting back. The root AI synthesizes everything and produces an answer.

No summarization. No deletion. No information loss. No decay. Every byte of the original document remains intact, accessible, and queryable for as long as you need it.

Now here are the numbers.

Standard frontier models on the hardest long-context reasoning benchmarks: scores near zero. Complete collapse. GPT-5 on a benchmark requiring it to track complex code history beyond 75,000 tokens — could not solve even 10% of problems.

RLMs on the same benchmarks: solved them. Dramatically. Double-digit percentage gains over every alternative approach. Successfully handling inputs up to 10 million tokens — 100 times beyond a model's native context window.

Cost per query: comparable to or cheaper than standard massive context calls.

Read that again. One hundred times the context. Better answers. Same price.

The timeline of the arms race makes this sting harder. GPT-3 in 2020: 4,000 tokens. GPT-4: 32,000. Claude 3: 200,000. Gemini: 1 million. Gemini 2: 2 million. Every generation, every company, billions of dollars spent, all betting on the same assumption.

More context equals better performance.

MIT just proved that assumption was wrong the entire time.

Not slightly wrong. Fundamentally wrong. The entire premise of the last five years of context window research — that the solution to AI memory was a bigger window — was the wrong answer to the wrong question.

The right question was never how much can you force an AI to hold in its head.

It was whether you could teach an AI to know where to look.

A human expert handed a 10,000-page archive does not read all 10,000 pages before answering your question. They navigate. They search. They find the relevant section, read it deeply, and synthesize the answer.

RLMs are the first AI architecture that works the same way.

The code is open source. On GitHub right now. Free. No license fees. No API costs. Drop it in as a replacement for your existing LLM API calls and your application does not even notice the difference — except that it suddenly works on inputs it used to fail on entirely.

Prime Intellect — one of the leading AI research labs in the space — has already called RLMs a major research focus and described what comes next: teaching models to manage their own context through reinforcement learning, enabling agents to solve tasks spanning not hours, but weeks and months.

The context window wars are over.

MIT won them by walking away from the battlefield.

Source: Zhang, Kraska, Khattab · MIT CSAIL · arXiv:2512.24601
Paper: https://t.co/ngovOSNrCQ
GitHub: https://t.co/gT0ootCNoa

[nicole grossmann @nicgrossmann]

As a mathematician who loves to study the brain, ML has never been so fun!

neurons (brain cells) 🤝 perceptrons (single layer neural net)

[nicole grossmann @nicgrossmann]

My GitHub just turned into a building. 68 contributions, Rank #56242. What does yours look like? https://t.co/bE4q1zJIxK

@samuelrizzondev stop destroying my building!!!! 🥺

[nicole grossmann @nicgrossmann]

Love stripe mentioning @firecrawl ! way to go @nickscamara_ 🇧🇷🇧🇷

[Patrick Collison @patrickc]

Just launched some great new providers in Stripe Projects (https://t.co/QwQF6d2TIz):

@huggingface
@Cloudflare
@OpenRouter
@firecrawl
@flydotio
@Amplitude_HQ
@mixpanel
@inngest

Many more coming later this month.

[Deedy @deedydas]

This Polish theoretical physicist just proved you can recreate all math functions from JUST one operation.

E(a, b) = e^a - ln(b)

Every single operation: +, -, x, / , trig, log, as you can see below.

Extremely mathematically elegant.

[nicole grossmann @nicgrossmann]

@firecrawl 🔥

[nicole grossmann @nicgrossmann]

@eglyman i wish people were also given telepathy. world kinda needs it

[nicole grossmann @nicgrossmann]

google's TurboQuant (ICLR 2026, arXiv:2504.19874) is making noise and honestly the core idea is almost offensive in its simplicity:

randomly rotate your vectors before quantizing. rotate back after.

that's it. 6x KV cache compression, 3-bit precision, 8x faster attention on H100s, zero retraining.

overrated hype or a real shift in how we think about inference optimization?

(also as someone who misses NYC: why are we still debating Euclidean vs Manhattan distance when Manhattan clearly wins in every dimension)

[nicole grossmann @nicgrossmann]

@nickscamara_ @firecrawl @brexHQ wow