Articles
Kyutai teaches a speech model to do arithmetic out loud
Kyutai released two open-weight speech-to-speech models that reason through math without transcribing audio first, posting a large jump on spoken GSM8K.
Editor's Picks
Hugging Face ships a $399 duck robot you can retrain
Hugging Face unveiled Microduck, a $399 open source duck robot that learns tricks through reinforcement…
Keenable scores $26M from Accel for an agent-first web index
A Yandex search veteran's startup raised $26M to build a web index tuned for AI…
ChatGPT Work signs into your web accounts all by itself
ChatGPT Work can now store login cookies in its cloud browser and sign into your…
NVIDIA doubles edge AI speed with Jetson Orin Nano 2 module
NVIDIA's Jetson Orin Nano 2 doubles entry-level edge inference and cuts power draw 40 percent.
Latest Articles
Google’s ToolGrad builds tool-use data by running the APIs first
A Google Research framework flips the usual recipe by executing a working API chain first, then writing the question that…
Sakana ships two Fugu models tuned to balance cost and quality
Sakana AI's new Fugu Max and Fugu Ultra v2 share one orchestration engine but optimize for different points on the…
AWS swallows DuckDB creator DuckLabs in database bet
The cloud giant is absorbing the team behind the wildly popular embedded analytics database while keeping it open source.
AWS opens a benchmark that tests agents on live cloud work
AWS's new open-source benchmark runs agents against real resources in disposable cloud accounts.
Developer survey paints AI coding as a habit that is hard to quit
A Coddy Tech survey finds 80% of developers feel their AI coding use borders on dependence.
Duke study finds AI creativity is converging across rival models
Duke researchers find creative output across 69 models has grown more similar over three years.
Microsoft’s AI chemistry model gets sharper and joins five simulators
Microsoft's Skala-1.1 model sharpens AI-driven chemistry simulation and lands in five major codes.
LinkedIn fields a team of AI reviewers that cross-check each other
LinkedIn's multi-agent review platform cross-checks suggestions across models and finds 63.9% get accepted.
