Nicholas Runcie (@NicholasRuncie) header
Nicholas Runcie (@NicholasRuncie) profile picture

Agents for Chemistry | Oxford DPhil student @OPIGlets | Opinions my own

Joined June 2022
Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)Photo by Nicholas Runcie (@NicholasRuncie)
Nicholas Runcie retweeted
At b12 Labs our mission is making the molecules AI designs. b12 Foundry is a significant step in that direction, and we’re excited to share that today! Synthesis is a complex problem of carefully orchestrated planning and execution. We have worked intensely on planning, but the execution is what makes it all real. We partnered with the HTEL at the University of Zurich to convert a state of the art automation facility into an advanced machine for producing molecules. At the b12 Foundry, our agents and chemists collaborate day and night to produce valuable molecules, reactions and data with a flexibility and quality that few other places in the world can demonstrate. Today, more than ever, it’s clear that the future of drug development starts with exquisitely designed molecules. We need to be able to make those complex molecules, and to make them efficiently, starting today. If you have molecules that are hard to make, need experimental data generated, or are simply excited by robots + chemistry + AI, get in touch. Read on in the comments 👇 #AIforScience #LabAutomation #Chemistry #DrugDiscovery #DeepTech #BioTech #AI
2
5
1
7
40,401
Nicholas Runcie retweeted
Very excited about this one: Today we're launching the b12 Foundry in partnership with the High-Throughput Experimentation Lab at the University of Zurich. AI can now design a molecule in seconds, but making it can still take a chemist several weeks. b12 Foundry closes that gap. b12-labs.com/foundry
14
18
3
58
4,741
Nicholas Runcie retweeted
At b12 Labs (YC S25) our mission is making the molecules AI designs. b12 Foundry is a significant step in that direction, and we’re excited to share that today! Synthesis is a complex problem of carefully orchestrated planning and execution. We have worked intensely on planning (link to selenium), but the execution is what makes it all real. We partnered with the HTEL at the University of Zurich to convert a state of the art automation facility into an advanced machine for producing molecules. At the b12 Foundry, our agents and chemists collaborate day and night to produce valuable molecules, reactions and data with a flexibility and quality that few other places in the world can demonstrate. Today, more than ever, it’s clear that the future of drug development starts with exquisitely designed molecules. We need to be able to make those complex molecules, and to make them efficiently, starting today. If you have molecules that are hard to make, need experimental data generated, or are simply excited by robots + chemistry + AI, get in touch. Read on in the comments 👇 #AI4Science #LabAutomation #Chemistry #DrugDiscovery #DeepTech #BioTech #AI
16
17
6
75
7,794
Nicholas Runcie retweeted
codex shipped the game while the car shipped the commute
1
2
195
Nicholas Runcie retweeted
If you want to make an exquisitely engineered molecule, as @maxjaderberg calls it, you need exquisitely engineered chemical synthesis to support it. One of the great-looking chemistry types that enable novel chemical spaces and are hard to automate is photocatalyzed reactions. Let's cook up a new generation of molecules together 👨‍🍳
1
1
3
448
Nicholas Runcie retweeted
Very excited to share our new OpenBind preprint, a more formal write-up of the first OpenBind data release and an extension of our earlier blog post. It adds more methodological detail, additional analyses, openly released outputs and code, and a new virtual screening benchmark.
6
43
3
197
23,680
Nicholas Runcie retweeted
Small personal update: after 4+ wonderful years at Cambridge, and a big detour working for the UK government, I'm beyond excited to share that I've joined the amazing team at @boltz_bio as a Member of Technical Staff, where I'll keep working at the intersection of AI + bio (1/5)
21
13
3
239
18,560
Nicholas Runcie retweeted
Excited to share our preprint: Molecular Representations for Large Language Models. We show that LLMs struggle with existing chemical formats, and that our new MolJSON representation substantially improves performance.
6
8
66
3,411
Nicholas Runcie retweeted
"Intelligence as a commodity" ~ NR #negroni #applepie
1
1
113
AIR Street Capital's "State of AI" 2025 Report references Nicholas Runcie, Charlotte Deane, and Fergus Imrie's work on Assessing the Chemical Intelligence of Large Language Models. Read the whole report here: stateof.ai/ And preprint here: lnkd.in/gbuNXR93
🪩The one and only @stateofai 2025 is live! 🪩 It’s been a monumental 12 months for AI. Our 8th annual report is the most comprehensive it's ever been, covering what you *need* to know about research, industry, politics, safety and our new usage data. My highlight reel:
1
1
4
1,229
GPT-5 sets new SOTA on ChemIQ - our chemical reasoning benchmark. 🧪💥⚗️ The model scored 70.2%, with a 14% jump from the previous best model. A thread on the chemical intelligence of GPT-5 🧵 @sama @gdb (1/n)
1
4
6
1,064
Really pleased to share what I have been working on for 2 months: 🇬🇧 UK SovAI are today announcing our £8m seed investment into OpenBind - A consortium that will actually make AI for drug discovery great by generating 500k experiment protein-ligand complexes!! Explainer 🧵 (1/n)
Excited to see this announcement from the Government’s SovereignAI unit: funding for the world’s largest dataset of protein interactions, led by a truly world-class group of researchers As @demishassabis says, “This is a brilliant initiative for UK science”
8
33
12
204
26,403
Nicholas Runcie retweeted
new paper from our work at Meta! **GPT-style language models memorize 3.6 bits per param** we compute capacity by measuring total bits memorized, using some theory from Shannon (1953) shockingly, the memorization-datasize curves look like this: ___________ / / (🧵)
83
395
55
3,567
637,093
🚀 LLMs can now do chemistry! Our new preprint shows that state-of-the-art reasoning models can now perform advanced chemical reasoning tasks, without any assistance from external tools. Here’s what o3-mini can already do👇 (1/🧵)
2
11
5
60
8,133
Replying to @gdb
NMR elucidation with o3 + RDKit is pretty cool.
2
8
1
28
3,500
Replying to @gdb
o3 can do full cheminformatics analysis from a single prompt. This is awesome!
1
2
1
12
2,245
Actually, it's only a novel hypothesis if the discovery comes from the hypothesé region of France. Otherwise, it's just a sparkling hypothesis .
4
5
92
5,954
Replying to @NicholasRuncie
Nice work! We were discussing this in our slack today. Matches our findings. One surprising thing for me is how much better latest Gemini is though - the may 7 one.
2
2
11
1,066