EN · International edition · 27/10/2026
Inferama daily · 27/10/2026
The most important AI developments since the previous newsletter, selected for editorial relevance.
News coverage period: 24/10/2026 08:30 → 27/10/2026 08:30 (Europe/Madrid).
OpenAI cuts frontier models off from live internet: the new offline training protocol
Following an 11-day pause triggered by security incidents, OpenAI has decided to eliminate real-time internet access for its most advanced models, replacing it with a simulated environment using cached web content.
Read article →
TAD improves a robotic policy in a simulated task without querying the teacher online
A preprint presents a method that learns from teacher-policy trajectories and raises OpenVLA-OFT’s success rate from 26.2% to 76.6% on one ManiSkill task. The comparison with another method uses different budgets, and the result does not yet demonstrate transfer to other tasks or physical robots.
Read article →
Audit reports lower rare-disease AI benchmark scores after stricter data isolation
A preprint on AI systems for rare-disease diagnosis reports that mean Recall@1 fell from 54.3% to 40.3% when strict separation between datasets was enforced. Its authors also report lower accuracy for supervised classifiers evaluated on external cohorts. The findings raise questions about how some benchmarks are measured, but do not by themselves establish these systems’ clinical usefulness or safety.
Read article →
RAQ-MLA tailors quantization to latent-attention caches and reports 2.5 bits per element
A preprint presents a quantization method designed for the two components of an MLA cache. Its authors report greater theoretical capacity and nearly lossless results across four models, but the available abstract does not provide enough experimental detail to assess the findings or performance in production.
Read article →
OpenAI Safety Researchers Dismissed: The Fight Over Confidentiality and AI Oversight
OpenAI fired three safety specialists for allegedly mishandling confidential information. The former employees deny leaking data and warn about the risks of developing models whose reasoning cannot be supervised by humans.
Read article →
AlphaIdeaBench evaluates research ideas against budgets and a 12-dimension rubric
The benchmark brings together 100 tasks with explicit resource limits and compares an AI judge with scores derived from review comments. Its results describe an evaluation method; they do not show that generated proposals are feasible in real-world projects.
Read article →
Block unveils Dessa: a native speech model that processes audio without intermediate transcription
Block has launched Dessa, an AI voice-to-voice model trained exclusively on human conversation audio. Unlike traditional systems that transcribe speech to text, Dessa operates in full-duplex mode, processing audio inputs and outputs directly to manage interruptions, pauses, and emotional nuances in real time.
Read article →