☀️ TRENDING AI NEWS
🤖 OpenAI launches GPT-Rosalind, its first life sciences AI model targeting drug discovery and genomics
🏢 Anthropic's Claude Mythos - deemed too powerful for public release - is heading to UK banks within days
🛠️ Physical Intelligence releases π0.7, a robot brain that can figure out tasks it was never explicitly taught
🚨 UK government makes its first investment under a new £500m sovereign AI fund
Something quietly shifted in the AI landscape this week - and it happened on three very different fronts simultaneously.
A specialized science model from OpenAI. A restricted-release AI tool landing inside British banks. A robot that improvises. None of these stories connect directly, but together they sketch out where AI is actually going in 2026 - deeper into institutions, deeper into the physical world, and deeper into fields that were once considered AI-proof. Let's get into it.
🤓 AI Trivia
What does GPT-Rosalind's name reference - the name OpenAI chose for its first life sciences AI model?
🧬 Rosalind Franklin, the chemist whose X-ray work was crucial to discovering DNA's structure
🔬 Rosalind Elsie Franklin, a fictional AI character from a 2019 novel
🧪 A NASA deep space telescope named after a botanist
🤖 An internal OpenAI codename that became the public name by accident
The answer is hiding near the bottom of today's newsletter... keep scrolling. 👇
🔬 OpenAI Enters the Drug Discovery Race with GPT-Rosalind
OpenAI has officially entered the life sciences arena. The company launched GPT-Rosalind, its first frontier reasoning model built specifically for biochemistry and genomic analysis - and the stated goal is ambitious: slash the 10-15 year timeline that drug discovery typically takes. This is OpenAI's first specialized science model, and it marks a meaningful departure from general-purpose AI.
From General Intelligence to Domain Expertise
The model is designed to work through advanced biochemistry problems and genomic data analysis - the kind of reasoning-heavy tasks that slow down early-stage drug research. If it works as advertised, this could compress years of lab iteration into months. It also puts OpenAI in direct competition with DeepMind's AlphaFold lineage and a growing number of biotech-focused AI startups.
This is a space worth watching closely. If you're tracking AI in healthcare and medical research, this is one of the bigger moves of the year so far.
🏦 Anthropic's Most Powerful Model Is Heading Into UK Banks - and Finance Leaders Are Nervous
Here's a sentence you don't read every day: a model deemed too dangerous for public release is about to be handed to British banks. Anthropic's Claude Mythos - which has so far been limited to a small group of primarily US firms - will expand to UK financial institutions within days, according to The Guardian.
Why the Finance Sector Is Both Excited and Worried
Senior finance figures are raising concerns even as their institutions prepare to onboard the tool. The tension is familiar: powerful AI capabilities that could transform financial services, but with risks that are genuinely hard to fully scope before deployment. Mythos has reportedly been restricted from wider release precisely because of its capabilities - which makes the decision to push it into banking environments a notable one.
We covered Anthropic's broader expansion push last week - the company is also plotting a major London office expansion with room to quadruple its 200-person UK headcount. The pieces are all moving in the same direction.
🤖 Physical Intelligence Builds a Robot Brain That Improvises
The holy grail of robotics has always been a general-purpose robot - one that can handle tasks it was never explicitly taught. Physical Intelligence just took a meaningful step toward that goal with π0.7, a new robot brain the company describes as an early but genuine move toward generalist robotic intelligence.
Tasks It Was Never Taught
The key claim here is significant: π0.7 can figure out how to handle tasks outside its training distribution. That's a meaningful distinction from robots that simply execute pre-programmed sequences. Physical Intelligence is positioning this as a step toward robots that adapt - not just repeat. For industries like warehousing, manufacturing, and home assistance, that flexibility is the whole game.
Physical Intelligence has been one of the hottest robotics startups to watch, and π0.7 suggests the research is moving faster than many expected. We discussed Google's Gemini Robotics advances earlier this week - the race for a general robot brain is now very much on.
🛍️ AI Shopping Traffic Is Up 393% - and Converts Better Than Regular Traffic
This number stopped me: AI-driven traffic to US retail sites jumped 393% in Q1 2026 compared to a year earlier, according to Adobe data. March alone saw a 269% spike. But the more interesting finding is what happens when that traffic lands - visitors arriving via AI tools convert better and generate more revenue than non-AI shoppers.
Intent-Driven Buyers Are Changing the Funnel
The likely explanation: people who've already asked an AI assistant about a product arrive at retail sites with much higher purchase intent. They've done their research via chatbot, decided what they want, and are ready to buy. For retailers, this is a meaningful shift in how AI is reshaping e-commerce - not just as a customer service tool, but as a top-of-funnel demand driver.
If you're building a retail presence or digital product, this is the stat to bookmark. And if you need a fast, AI-built website to capture that traffic, 60sec.site lets you spin one up in - you guessed it - about 60 seconds. Worth a look if you haven't already.
🏛️ The UK Puts £500m Behind Homegrown AI - and Makes Its First Bet
The UK government is backing its AI ambitions with real money. Technology Secretary Liz Kendall announced the first investment under the country's £500m sovereign AI fund yesterday, taking a stake in a British AI startup. The move is explicitly designed to reduce UK dependence on US and Chinese AI infrastructure.
'Seize the Opportunity' - or Get Left Behind
Kendall played down concerns about job displacement and cybersecurity risks, urging the public to "make AI work for Britain." The timing is notable: Anthropic is simultaneously expanding its London footprint, and the UK is clearly positioning itself as the European AI hub of choice. Whether the £500m fund is enough to compete with US-scale investment is a legitimate question - but at least the intent is there.
Wired also has solid coverage on the broader UK sovereign AI push and what it means for the country's tech independence strategy.
🌎 Trivia Reveal
The answer is A - Rosalind Franklin! The pioneering British chemist's X-ray crystallography work in the early 1950s was foundational to understanding DNA's double helix structure. It's a fitting name for an AI model built to accelerate the kind of biological research Franklin spent her life on.
💬 Quick Question
Would you feel comfortable knowing your bank was using a powerful AI model that was considered too risky to release to the general public? Hit reply and let me know your gut reaction - I read every response and genuinely want to know where you land on this one.
That's it for today. For more daily AI coverage, head over to dailyinference.com - and we'll see you tomorrow with more from the week that keeps delivering. 🤖