Living within the experiment
London, UK
Joined May 2026
- Tweets24
- Following12
- Followers5.3K
- Likes77
Pinned Tweet
1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵
Last week, we called on the Government to scrap unfair labour laws that hold back the UK’s AI sector.
Today, the Prime Minister announced plans to end the abuse of non-competes and go further on garden leave.
Credit to the Government and @KanishkaNarayan for moving at startup speed!
cityam.com/burnham-to-clamp-…
Much of scientific research runs on old software with clunky interfaces or unexplained bugs. To realise the benefits of AI for science, we first need to fix the foundations.
Join us to pay off science’s tech debt. Tokens, travel and accommodation provided, prizes to follow...!
London AI × Science Hackathon 👾
Science Works × @inherent_labs
• Oct 16-18
Lots of scientific software is outdated, hard to use, and isn't interoperable – let alone making use of (or ready to interface with) AI.
Or it's expensive. Join us to build better for the community
Some of the UK's leading founders and investors have signed our open letter since we launched this week.
Thanks to everyone who has joined the campaign to give talent freedom to move and make Britain the world's most innovative economy.
Add your name: freetostartandscale.uk.
Free to Start and Scale
An open letter to the UK Government. 29 September 2026.
freetostartandscale.uk
1/ Today we’re launching an open letter calling on the government to end unfair restrictions that stop Britain’s best switching jobs or starting companies.
UK AI companies that have raised $5bn agree: our best should be free to start and scale. 🧵
@downingstreet @biztradegovuk
Inherent retweeted
London AI × Science Hackathon 👾
Science Works × @inherent_labs
• Oct 16-18
Lots of scientific software is outdated, hard to use, and isn't interoperable – let alone making use of (or ready to interface with) AI.
Or it's expensive. Join us to build better for the community
1/ Today we’re launching an open letter calling on the government to end unfair restrictions that stop Britain’s best switching jobs or starting companies.
UK AI companies that have raised $5bn agree: our best should be free to start and scale. 🧵
@downingstreet @biztradegovuk
3/ At Inherent, we’re cutting non-competes and garden leave from our contracts entirely. This brings us closer to the California standard.
We'll keep our team because they believe in our mission of discovering new knowledge, not because we made it hard to leave.
freetostartandscale.uk
/4 You can sign the open letter using the link, and read more in this morning's @FT.
ft.com/content/930465de-0cf3…
Our CTO @kallyaleksiev is speaking at London Systems Club, sharing some infrastructure lessons from training Faraday.
Come join us on October 12 for a conversation about how to engineer scientific intuition.
Next Systems club on the 12th of October. This is a room of 20-30 of the best engineers in London that care about systems, hardware, and performance.
Speakers ⬇️
- @kallyaleksiev from @inherent_labs - Presenting lessons from Faraday, the infrastructure built to train scientific intuition on top of coding agents and outperform leading models at research paper replication.
- @ismaeel_bashir_ from @ExpanseCompute - 538,000 production jobs on a national supercomputer say your users are over-reserving by 2–3x. Here's how to predict what they'll really need.
Link to register in the comments 🔗
Nice to see some of our work covered in @nytimes! A bit of additional context:
A big piece of what motivated us to found Inherent was a belief that the use of AI for R&D requires new infrastructure and workflows that enhance human control. We are building tools for human-machine teaming that aid monitoring and oversight as models become more powerful.
These capabilities enable recursive self-improvement at the organisational level: a virtuous cycle that increases the collective intelligence of humans and machines. We see this as both a better outcome and a safer trajectory than model-only RSI.
We will share more detail soon about what this looks like in practice and how we approach safety and societal benefit. In the meantime, you can find our first research paper here: inherentlabs.ai/research/tra… and our manifesto here: inherentlabs.ai/
Reliable evaluations and responsible pacing are critical to developing AI that is safe and trustworthy. Success will require coordination across industry, including with newer labs.
At Inherent, we’re building systems for human-machine teaming that improve monitorability, increase human agency, and enable societally beneficial forms of cooperation. We look forward to playing our part.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Our very own @edwardfhughes joined MLST to talk about the nature of creativity, what the Tristan chord tells us about AI today, and our mission to discover new knowledge.
We spoke with @edwardfhughes from @inherent_labs about creativity and automating scientific research!
8/ We are designing Inherent so Faraday’s improvements compound through the entire company, enabling us to discover new knowledge while firmly keeping humans in the loop. To join us on our mission, apply here: inherentlabs.ai/.
Paper: arxiv.org/abs/2608.13331.
Blogpost: inherentlabs.ai/research/tra….
