adaption

Value comes from AI that keeps learning inside the business, not from the biggest model.

@sarahookr at the @ftlive Future of AI summit with Ken Moore (@Mastercard), Philipp Heltewig (@cognigy), Nathan Marsh (@Bentley Systems), and @andrewtghill.

Adaption is now a plugin in @ChatGPT and Codex.

Describe the data you wish you had, and Adaption invents it. Already have data? Adaptive Data gets it ready for training, and AutoScientist adjusts the recipe to fit your custom model.

Better models start where your work already happens.

Can the Adaption team name a moment in computer science history from one tiny detail?

Co-founder @sudip_r0y and Scientific Lead @hugo_larochelle were put to the test with the team.

New Adaption research: with AutoScientist, a lower-cost model outperformed GPT-6 Astra and Gemini 4 Argon on long medical records, at up to 15.4x lower cost.

Correct answers jumped from 58% to 75%. No retraining.

The Adaption team rates AI’s biggest debates from 1 to 10.

@sarahookr, @sudip_r0y, @hugo_larochelle, and @weiyinko_ml gave open weights a unanimous 10. Parallel parking did not fare as well.

What’s your score for AI agents?