Tom Hulme (@thulme) header
Tom Hulme (@thulme) profile picture

Managing Partner @GVTeam (Google Ventures) https://xitter.gq/t.co/HO3pTc6UVa // MBE for Military Service in 2025 // https://xitter.gq/t.co/aFGhfyemAm Previously founded @OpenIDEO @IDEO

London
Joined July 2007
Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)Photo by Tom Hulme (@thulme)
Lunch with @guypod improved my list of where frontier AI still isn't enough. Cyber and funds are adversarial: the other side gets the same models, so the bar keeps moving. Science is vastly complex, but nature isn't playing against us, so that demand should flatten.
2
5
422
Most tasks have a finish line, and the frontier models are past it. In knowledge work, three fields stand out for having none: Cyber security, where the bar is set by an adversary with the same models. Science and drug discovery, where it's set by nature. Long-horizon investing, where consensus is already in the price. No coincidence that frontier labs now lead flagship launches with cyber and science.
2
1
7
806
Thrilled to announce $5B in funding for @IsomorphicLabs to accelerate our mission to Solve All Disease. We continue to make rapid progress reimagining drug design from first principles with AI to radically cut the time & cost to find new medicines. Gradually, then all at once...
639
1,591
416
14,784
1,012,930
In 1973, GCHQ mathematician Clifford Cocks invented RSA public-key encryption four years before MIT published it, then kept it classified for 24 years until 1997, forfeiting every patent! That feels like a different world!
3
806
Break most office jobs into individual tasks and frontier models can already do the bulk of them. Where rollouts actually stall is the joins between tasks: permissions, messy context, handovers, and smart people simply not yet realising what's possible.
1
6
930
Knowledge work has moved from the 80/20 rule to the 99/1 rule: AI hands you something 99% finished in a minute. Doing 1% of the work sounds restful. Instead, a March HBR/BCG study coined “AI brain fry”: workers pushed into heavy AI oversight report 33% more decision fatigue and 39% more major errors, with strain surging past 3 concurrent AI tools. Why? Because the old 80% slog (formatting decks, tidying models) was disguised recovery where your brain worked out the argument. Delete the cruise, and the whole day is take-offs and landings.
40
115
40
1,235
236,926
The history of drug discovery is a slow battle against biological entropy. @IsomorphicLabs is changing the clock speed, moving toward experimental accuracy and making meaningful progress in its mission to solve all disease. GV is excited to continue backing this world-class, London-built team again in their Series C! Congrats @demishassabis @maxjaderberg and team 🚀
We’ve closed our $5 Billion funding round co-led by Greenoaks and Sequoia Capital to build the future of medicine and bioengineering with AI. isomorphiclabs.com/articles/…
5
4
58
2,358
Tom Hulme retweeted
Last week we hosted @RichardSocher (Recursive) and GV's @thulme in SF for a fireside with researchers and founders from Google DeepMind, OpenAI, Mistral, Reflection AI and more. Sharp room, sharper questions. A few takeaways 🧵
1
3
18
4,170
We are notoriously bad at feeling the pace of progress when we're inside it Last week at @GVteam SF, I hosted @RichardSocher and a brilliant room of researchers from GDM, OpenAI & Mistral. We talked through his awesome new book, Eureka Machine, & the amazing progress @Recursive_SI 🧵
2
2
14
1,178
Some of his thoughts that stuck with me: Timeline compression is real, and the future needs better marketing. When we were born, machines were worse at almost everything. In our lifetimes, they’ll beat us at almost everything intellectual. More science → more technology → more growth → more human flourishing
1
3
333
The proof is in how we now build AI. @Recursive_SI just built Flint, a frontier-class model, in under 4 months with a tiny team. How? The Eureka Machine automated the research. When the machine starts building the machine, progress compounding is no longer theoretical.
2
182
AI agents will soon move money on our behalf, for that to work, the system needs to: know who is transacting, follow the rules, remain secure, keep data private and do it all at speed and low transaction cost globally That's what @bzlwang @cclayporter @awxjack and the @altaxyz team are building We @GVteam are excited to join their Series A
At Alta, our mission is to optimize the global financial system with frontier blockchain infrastructure. And today we’re announcing $80M in funding, including our Series A led by @IndexVentures with participation from @GVteam, @blackbirdvc, @SquarePegCap, and Capital 49.
3
1
32
1,719
Tom Hulme retweeted
Across our projects @GoogleDeepMind, we are witnessing a genuine phase transition in the pace of scientific discovery. You will see a glimpse of this in our releases in the next few weeks.
300
785
286
9,371
1,191,858
Tom Hulme retweeted
Great to see @Recursive_SI in the top 10 in talent movement. We are actively hiring: recursive.com/join-us Join a fantastic team of ex OpenAI, Anthropic, DeepMind, and Meta members of technical staff.
Top 30 Startups and BigTech companies based purely on ~50,000 engineer job changes in the last 3 years. If engineers leave a company to join another, it ranks higher especially if the company they come from ranks highly. Here are the final results:
5
4
2
89
16,983
Tom Hulme retweeted
110
126
87
1,002
159,934
Tom Hulme retweeted
Software Factories are extremely valuable, but hard to build. @tessl_io's new product may have just solved that 🙂 - Simple onboarding turns a repo into a factory in moments. - Template loops for code review, vulnerability remediation, cost optimization and more. - Built-in self-improvement for loops, compounding your factory’s value. - Interactive factory floor for tracking activity and solving problems. - Clear and simple metrics to know what you’re getting. Tessl factories are fully built on open skills, which you create, hone and own indefinitely. We help you create skills from your history, and the factory starts to self-improve them from here. Lastly, Tessl factories are declarative and easy to understand and manage. Config files composing skills and environments into actions, actions and triggers into loops, and multiple loops into a factory. This is the equivalent of Infrastructure-as-Code for factories. Instead of a bunch of disparate scripts you piece together to understand what’s happening, they codify what the factory does. This means Tessl factories aren’t just easy to setup, they’re easy to manage too. At the end of the day, our job as tool builders is to delight our users. Solve real problems with strong technology and great UX, that help them run fast. I feel the Tessl Factory is getting this job done. If it wasn’t obvious, I’m excited by our new product 🙂 At the end of the day, our job is delight our users, and I firmly feel the Tessl Factory does so. As part of the launch, we’re offering $200 free credits to all users. Try it out, and let us know what you think - I can’t wait to see what you build.
21
19
6
216
25,019
Tom Hulme retweeted
We now serve almost 50% of the Fortune 50, and their ambition is driving our growth. Today, we hit $400M+ in ARR, adding over $150M+ in ARR in Q3 alone - that's more in one quarter than we generated all of last year. Also, our growth is accelerating: we took 6 months to double from $100M to $200M ARR, and only 5 months to double from $200M to $400M. Read more: sierra.ai/blog/our-customers…
6
7
4
177
32,117
This "messy middle" is now hitting everyday services too, quietly breaking legal claims, insurance, ombudsmen, and customer support. For decades, organisations relied on human effort as a natural rate-limiter. Writing a formal legal claim, regulatory complaint, or detailed appeal took hours of work or hundreds of pounds in legal fees. That friction kept inbound volume matched to the humans sitting on the other side of the desk reading them. Generative AI just dropped the cost of creating those claims to zero. Today, anyone can generate a 20-page legal claim or dispute letter in thirty seconds. Look at the UK Legal Ombudsman's latest figures: more than half (51%) of complaint files now show clear signs of AI drafting, helping drive a 37% surge in annual caseloads. Because most receiving organisations still rely on humans to read and triage every submission, zero-friction AI generation acts like a live DDoS attack on their teams. You cannot solve a machine-speed flood by hiring more people to read PDFs. The only way through this messy middle is AI customer service on the receiving side too, triaging, verifying, and resolving issues at machine speed. That is a big part of why I am so excited about our investment in @SierraPlatform (@btaylor and @claybavor).
Self-driving cars and AI security share the same messy transition: mixing humans and machines. On roads, mixing unpredictable humans with AVs creates friction before all-AV fleets make safety soar. Online, AI attack vs human defence is broken. We need AI on defence.
3
1
7
1,204
Self-driving cars and AI security share the same messy transition: mixing humans and machines. On roads, mixing unpredictable humans with AVs creates friction before all-AV fleets make safety soar. Online, AI attack vs human defence is broken. We need AI on defence.
1
6
1,751