People Behind AI, San Francisco edition.
Yesterday we invited people working in AI to the Hilton rooftop for a cocktails & food evening after @COLM_conf
Want to join us at the next one? Follow us and stay tuned. Thanks to everyone who joined us.
San Francisco we are here!
@COLM_conf starts today, and the Sovrano team is officially here. If you’re around, come meet us! We’ve got a few things happening at the booth over the next couple of days, so come find us and say hi :)
1,087 papers in, 16 out.
This week: DPO was never built for how language models actually generate text; and three separate teams just found the same blind spot.
The research worth reading →
https://t.co/kl58oYQNqq
AI sovereignty is becoming an enterprise priority.
Trust, control and visibility over AI systems are no longer just technical concerns, they’re strategic ones ↓ https://t.co/fejBcFYKwU
Next stop for Sovrano: @COLM_conf 2026.
We’re heading to San Francisco to spend a week with researchers and teams working at the frontier of AI.
Find @SovranoAI at our booth, or message us to grab a coffee ;)
We’re travelling 9,500 km from Barcelona to SF to attend @COLM_conf 2026 next week.
We’ll have our own booth, a few activities to put your AI judgement to the test, and some gifts waiting for you!
And if you’d like to grab a coffee and chat during the conference contact via DM ☕
See you in SF!
WE'RE HIRING!!
3 roles open in Barcelona:
→ Project Lead: https://t.co/AP09qNot0l
→ Talent Ops & Community Associate: https://t.co/7adSMns3pm
→ University Partnerships Associate: https://t.co/CpIMPYsY19
We’re moving fast, working on ambitious AI projects, and looking for people who want to take real ownership as we grow. We want to know you!!
the best conversations happen after the talks, around a table
so we're doing that on purpose. People behind AI, by @SovranoAI
madrid, this thursday. limited seats.
building or researching AI in madrid and want one? dms open
When I started Sovrano AI, the idea was simple:
What if people with real expertise had a bigger role in how AI gets built?
A lot has happened since then. So with many new people here, it feels like a good time to go back to the beginning.
This is Sovrano AI 👀 let's make Europe sovereign 💡
People Behind AI is coming to Madrid!!! 🥘
Good food. Good conversations. And a table full of people shaping AI.
Want a seat at the table? DM @marinaalvaro7 🤝
930 papers in, 15 out.
This week: one uncomfortable finding: your benchmark may be measuring your annotators
The research worth reading →
https://t.co/4wa4nWxYtn
What does “better model performance” actually mean?
In EuroExec, 6 frontier models were evaluated across 413 real-world executive questions by 47 domain experts using rubrics, checklists and preference ranking.
The leaderboard is the output. The evaluation methodology is what gives it meaning.