Today, @bayesiankalki and I are launching Mnemonic Labs (YC F26). We're a neolab training a foundation model for quant trading that learns directly from market activity. In <6 weeks we’ve built our training pipeline and model architecture. Our current training run uses over 25TB of tokens.
We met 28 years ago shortly after I was born (we’re brothers!). We’re Olympiad finalists (5x USACO combined) and MIT graduates. Kalki made $20M in annual PnL at 22 at Tower Research and I was the youngest quant at Citadel at 19. Over the past ten years, we’ve been part of quant trading’s evolution from linear models to trees and now neural networks.
Learn more about us at ycombinator.com/launches/XRK… or reach out at [email protected]!
Videh Seksaria retweeted
Today, we are excited to introduce Ladon, a state-of-the-art biological foundation model that predicts disease trajectory and treatment response in patients. There is immense complexity in the individual biology of every patient. Two patients can have the same diagnosis and receive the same treatment, yet have very different outcomes: one responds well and lives for years, while the other’s disease continues to progress. Understanding and resolving this heterogeneity to predict outcomes is how we accelerate our delivery of medicines for patients and unlock clinical trial abundance.
Ladon is our first step towards this future: one where we can learn more from every patient enrolled so that we can run smaller, faster trials with higher success rates. We built it to give drug developers quantitative support on billion-dollar clinical trial decisions. Today we're excited to share Ladon's capabilities in our blog and technical report.
Ladon is a 3B parameter model trained on millions of RNA-seq samples spanning decades of technical and biological variation across tissues and species. Ladon beats gene expression baselines, established biomarkers, and existing RNA foundation models at predicting patient prognosis and therapeutic response in 140+ oncology tasks, including external validation on past clinical trials. This is significant because clinical trials are fundamentally generalization problems. We show that Ladon-derived predictions are robust across indications, treatments, and sequencing platforms.
Reach out if you’d like to get started with the model today!
Read more about Ladon in the blog: blank.bio/introducing-ladon
Find more technical details in the white paper: blank.bio/introducing-ladon
Can’t imagine better founder market for!
We got into @ycombinator to fix how AI makes music.
Today we're launching Taan v1-preview, the most controllable foundation music model.
Music models today make catchy tracks, but you can't steer them. Ask for "violin at exactly the 11th second" and they usually skip it. So you regenerate over and over and hope one take fits.
Taan is trained to fix this.
Hum a melody, name exact notes, or hand it an image, a MIDI file, a video or a script. Taan composes music that fits your context and follows your instructions.
v1-preview has genre support limited to orchestral and cinematic music, with more genres on the way.
live today 👇
6
3,995
whoa this is such a cool video. and ofc an incredible team and opportunity!
Epsilab is working with frontier labs to build RL environments for long-horizon enterprise work.
Today, low-quality data is an industry-wide problem. We founded Epsilab to bring together real-life data and top 1% domain expertise to solve this.
Backed by @ycombinator (F26).
Epsilab Launch
YC F26. Frontier intelligent for RL
5
1,187
whoa this is going to sell like hotcakes
The world is borrowing hundreds of billions of dollars to build GPUs, and the average GPU is busy 5% of the time.
Nodus Compute gets more AI out of the GPUs that already exist.
We fill idle hours with useful work, pack more models and jobs onto every card, and run training, RL, agents and inference on one platform.
3
424
in the hands of the wrong practitioner, the bitter lesson can be used to justify whatever misguided choice the research wants
it's hilarious how many have taken the bitter lesson to mean just scaling up transformers. it's about scaling, but innovations in algorithmic primitives can lead to much more efficient scaling
3
457
This is super cool!
I was shocked that 3/4s of designers work in coding agents rather than design tools. This is crazy?? Designers deserve a tool built for them, not just for developers.
That's why I'm building @proto_ide, the designer’s IDE. The shape of Claude Code and Cursor, but with every design feature you wish they’d ship.
Early access signups are open. If you want a tool that finally unlocks your full creative potential, sign up at prototypes.fun :)
2
6
723
the old pinterest trick! love to see it
Check out graypass.com on all the MacBooks in Apple Store grand central New York
1
3
414
These guys are awesome!
Today, in the first of a series of blog posts we’re calling “Behind the Grove”, we’re offering a look into some of the technical work we do at DeepGrove.
Our first post is about our pretraining infrastructure, Grove. Grove is built around 2 backends: Grove-FSDP and DDP. We detail how we optimized both backends, culminating in an over 2x speedup over public baselines.
A production-scale pretraining run of Maple 20B-A1B (deepgrove.ai/maple-preview) on 4T tokens takes a mere 2 weeks on 64 H100s.
We view infrastructure as the backbone of our work, and as one of the many pieces that help us build toward the frontier of efficiency and performance.
To read the full post, visit deepgrove.ai/infra.
3
448
This is so cool! 600 tokens/second is a game changer
DeepSeek V4.1 Flash 1017 runs at 600+ tok/s on Isoquant
Ships Tetris in 10 seconds, 8.7x faster than Claude Opus 5.5.
This is our AI optimizing AI, and we are passing performance wins and cost savings on to our customers:
→ 0.24s P50 / 1.16s P95 time to first token
→ $0.035 input / $0.005 cached / $0.37 output per 1M tokens
Keep your code and harness, just change the base URL. Works with your existing OpenAI and Anthropic SDK integrations.
No waitlist, live now: isoquant [dot] ai
This is insanely cool!
Live out your wildest fantasies on Yam (YC F26): AI-powered romance simulation games for women, built by women 💃
Walk into a dating villa as the new bombshell. Star in your own K-drama. Fall in love with your enemy turned lover. Describe a world and who you want to be, and Yam brings it to life.
- Beta players have spent collectively ~1 million hours on our platform
- Yam girlies play on average 3 hrs/day.
Build your world on yam hiyam.ai/
4
341
I hate san francisco halloween because what do you mean you are giving out copies of the commanding heights
2
330
what is it about graph neural networks that makes them so enticing to researchers yet never work?
1
7
517
finally! someone implemented Schmidhuber's Neural Sequence Chunkers (1991)
I've always believed that intelligence should come from simple, general building blocks that learn their own abstractions directly from raw data; in other words, the bitter lesson. I think we're finally seeing a real step towards AI that can learn from all data.
Language modeling on raw bytes is a microcosm of this ideal, and to me it's a central problem in AI: if we can't build LLMs that learn end-to-end from bytes, how will we handle the rest of the world's data that we understand even less?
Last year we introduced hierarchical networks (H-Nets) as a first step. Our new version finds up to 3-5x efficiency gains in controlled scaling law experiments, and at larger scale matches an open-source model trained on 11T tokens with ~10x less data and compute. The posts below walk through what we found and why we think it works.
3
391

















