Lightnews — Scholar-powered news

Andrew White 🐦‍⬛ @andrew.diffuse.one · 12d

So we probably won't be getting a direct simulation of a whole virtual cell at meaningful timescales any time soon. Oh, and it would require 20x current earth power generation. 3/3

Read the analysis/blog post here: diffuse.one/p/d1-009

diffuse.one

andrew white's blog.

diffuse.one

3

Andrew White 🐦‍⬛ @andrew.diffuse.one · 12d

It sounds insane, but remember there are 10^14 atoms in a human cell and 10^20 femtoseconds in a day. And across multiple simulation engines, it requires 10^4 FLOPs per atom x femtosecond 2/3

1 1

Andrew White 🐦‍⬛ @andrew.diffuse.one · 12d

I finished my estimate on required compute to make an atomic-resolution virtual cell: 10^38 FLOPs to simulate a human cell for 1 day. We should be able to do this simulation in 2074 using 200 TW of power. 1/3

4 3 11

Andrew White 🐦‍⬛ @andrew.diffuse.one · 19d

yea, those are the model thoughts. It has a lot of mistakes in its thoughts. But you've got a very good eye! We'll make sure the final paper has a pristine example of its thoughts.

1

Andrew White 🐦‍⬛ @andrew.diffuse.one · 19d

Our ether0 paper was accepted at NeurIPS 2025! Very proud of the FutureHouse team!

1 6

Andrew White 🐦‍⬛ @andrew.diffuse.one · 22d

Very good point - I can re-run without that phrase.

1

Andrew White 🐦‍⬛ @andrew.diffuse.one · 22d

If you don't put phrase in quotes, it's an or. So it was

"α" equation

which is equivalent to "α" OR equation

1

Andrew White 🐦‍⬛ @andrew.diffuse.one · 24d

You can also look at it over time. Here's relatively popularity of different animal models in research over time.

Anyway, found this to be interesting. More details about it here: diffuse.one/p/d2-003 3/3

1 3

Andrew White 🐦‍⬛ @andrew.diffuse.one · 24d

Here's one measuring the frequency of sample sizes. Like how often people use 8 samples vs 12 samples for reporting research results. N=2 is apparently the most popular 2/3

2 3

Andrew White 🐦‍⬛ @andrew.diffuse.one · 24d

Google scholar has a full-text index of nearly all research papers. You can use it to get counts for arbitrary phrases. I've been using this to measure popularity of things in science. For example, here's the popularity of Greek letters used in equations 1/3

3 1 13

Andrew White 🐦‍⬛ @andrew.diffuse.one · Aug 15

read it here: diffuse.one/p/d2-002

diffuse.one

andrew white's blog.

diffuse.one

1

Andrew White 🐦‍⬛ @andrew.diffuse.one · Aug 15

I've written up some thoughts on publishing for machines. 10M research papers are published per year and there are 227M total - machines will be primary producers and readers of publications going forward. Humans can simply not keep up. It's time to think about revising the scientific paper.

1 1

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 23

We make evals at FutureHouse. It’s hard and it sucks. It’s also now the bottleneck, as we scratch the boundary of human ability. HLE was a huge effort and made many good questions and we hope this analysis stimulates review of the other HLE categories and improvements 7/7

1 2

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 23

We have written up our analysis: www.futurehouse.org/research-ann...
And made a gold subset on @huggingface that passed our review: huggingface.co/datasets/fut... 6/7

futurehouse/hle-gold-bio-chem · Datasets at Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

huggingface.co

1 2

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 23

We reviewed 150 of the questions in the chem and bio and found about 30% have peer-reviewed papers contradicting their ground-truth answers. Issues include confusion of species with orders, misreading of FDA guidelines, etc. All our notes are public. 5/7

1

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 23

The HLE rubric wanted questions to have “objectively correct, univocal” ground-truth answers. You can find multiple peer-reviewed papers that contradict the statement "Oganesson was the rarest noble gas in 2002 as a percentage of terrestrial matter" 4/7

1

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 23

It’s a clever question. But it’s not really about frontier science. Multiple papers have shown that Oganesson is not a gas (it’s predicted to be semiconducting solid), it’s not noble (it’s reactive), and it isn’t included in any "terrestrial matter" tables of noble gases. 3/7

1

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 23

The design process of HLE required the questions to be unanswerable by contemporary LLMs. That lead to many gotcha style questions like the one below. It’s a trick question – in 2002, a few atoms of a group 18 element Oganesson were made for a few milliseconds. 2/7

1 1

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 23

HLE has recently become the benchmark to beat for frontier agents. We at FutureHouse took a closer look at the chem and bio questions and found about 30% of them are likely invalid based on our analysis and third-party PhD evaluations. 1/7

1 2 6

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jul 12

I just noticed it has sound lol. It's amazing

Reposted by Andrew White 🐦‍⬛

The Align Foundation @alignbio.bsky.social · Jul 8

1/4
🚀 Announcing the 2025 Protein Engineering Tournament.

This year’s challenge: design PETase enzymes, which degrade the type of plastic in bottles. Can AI-guided protein design help solve the climate crisis? Let’s find out! ⬇️

#AIforBiology #ClimateTech #ProteinEngineering #OpenScience

1 20 23

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jun 22

It may take a bit to extract the function, but here it is: github.com/Future-House...

ether0/src/ether0/rewards.py at c8cc676354e926b50ad206a606e04489bc9c95e3 · Future-House/ether0

A scientific reasoning model, dataset, and reward functions for chemistry. - Future-House/ether0

github.com

2

Andrew White 🐦‍⬛ @andrew.diffuse.one · Jun 22

I have written up a 3.5k word/10 figure essay on how to write a reward function while avoiding reward hacking for chemistry. It covers all the ridiculous ways we had to avoid reward hacking for training ether0, our scientific reasoning model.

diffuse.one/p/m1-000

2 1 23

Andrew White 🐦‍⬛ @andrew.diffuse.one · May 20

Although the discovery here is exciting, we are not claiming that we have cured dry AMD. Fully validating this hypothesis as a treatment for dry AMD will take human trials, which will take much longer.

Blog: www.futurehouse.org/research-ann...
Paper: arxiv.org/abs/2505.13400

Demonstrating end-to-end scientific discovery with Robin: a multi-agent system | FutureHouse

www.futurehouse.org

4

Andrew White 🐦‍⬛ @andrew.diffuse.one · May 20

The code for this is really minimal - similar to Google Co-Scientist we used multiple agents (from our platform in this case) and tournament-style rankings to select ideas. We're open sourcing it next week, along with all the trajectories.

1 2