Agents in production (2024): tool use with guardrails
Enterprises deployed agents with real tools—CRM, code, browser—under tracing, permissions, and continuous eval.
Sorted by date · America/Guayaquil timezone
Latest signals
Page 32 / 38 · 447 articles
Realtime voice agents: the spoken interface returns
Full-duplex realtime audio APIs enabled low-latency call centers and spoken assistants with tool use.
California’s SB 1047: the U.S. debate on risk thresholds
California’s proposed frontier-model safety duties split labs, academia, and Silicon Valley before the veto.
OpenAI o1 (2024): models trained to reason longer
The o1 series introduced inference with internal “chain of thought” optimized by RL, excelling at math and code.
AI Act takes effect (2024): from agreement to obligation
The European regulation began compliance timelines for bans, transparency, and general-purpose models.
Grok and xAI: another frontier lab in the mix
Elon Musk’s xAI launched Grok inside X, betting on a distinct personality and realtime social data.
AI at math olympiads: medals and verification
DeepMind and OpenAI systems reached medal-level IMO performance with search and formalization pipelines.
AlphaProof and AlphaGeometry: formalization as muscle
DeepMind coupled language models to Lean and synthetic geometry to solve olympiad problems with checkable proofs.
Energy and data centers: AI’s physical cost
Electric demand from training and inference clusters strained grids, PPAs, and climate agendas.
Claude 3.5 Sonnet: the practical leap in software engineering
Anthropic positioned Claude 3.5 Sonnet as a favorite for coding and dev agents on quality and cost.
Apple Intelligence (2024): on-device AI and private cloud
Apple announced a hybrid stack: on-device models and Private Cloud Compute for heavier tasks, with a privacy focus.
Evals and red teaming: industry professionalizes measurement
Labs and governments expanded batteries testing dangerous capabilities, jailbreaks, and reliability before deploy.