#Trend

AIPublic AI benchmarks such as MMLU, SWE-bench Verified, and GPQA Diamond are saturating or showing contamination, prompting OpenAI to call SWE-bench Verified "no longer suitable" in late February and recommend SWE-bench Pro. OpenAI's audit found 59.4% of the problems its best model failed had flawed test cases, and GPT-5.2, Claude Opus 4.5, and Gemini 3 Flash could reproduce original fixes from memory. The article argues that behavioral tests, such as Vending-Bench's simulated vending machine business, may be more useful for everyday model choice.

AIChris Olah, identified as an Anthropic account, shared a post pointing to coverage by Kevin Frazier, a national security law professor, with Alan Rozenshtein. The post itself gives no details about the topic or findings covered.

AIChatGPT has crossed 900 million weekly users and 50 million paying subscribers, according to OpenAI's Nick Turley. He says people use it for writing, building, research, trip planning, shopping, and everyday tasks, and that growing usage is helping improve response speed, reliability, and naturalness.

AIMeta's Ahmad Al-Dahle says some of the most talented engineers he knows are quietly panicking about their careers. He argues they shouldn't, while noting the profession is changing faster than most people realize. He links to a Medium article covering what is shifting and what matters next.

AIAI Futures Project grades its AI 2027 scenario for 2025 and finds quantitative progress running at roughly 65% of the predicted pace, later revised to about 75%. Most qualitative predictions, such as the rise of coding agents, are judged on pace, while SWE-bench-Verified progress was slower than forecast and OpenAI's valuation trailed the scenario. The authors say their timelines lengthened over 2025 and plan to keep updating forecasts through 2026.

AIAI researcher Chip Huyen asks readers how long they think it will take AI to fully automate their jobs. The post is a question with no data, benchmark, or stated prediction.

AIYi Tay announced that a new frontier lab in Singapore, described as the first in Asia, is building AGI and gathering top talent. He welcomed @zzlccc, who joined Google DeepMind to work on Gemini post-training under Tay's leadership.

AIGeoffrey Hinton called the International AI Safety Report 2026 a thoughtful, detailed, and well-researched description of AI risks, essential reading for anyone writing or speaking about them. Yoshua Bengio's thread introduces the report as the most comprehensive evidence-based assessment of AI capabilities, emerging risks, and safety measures to date.

AIYi Tay, who recently hired a full AI team from thousands of applications, says PhDs are still a reasonable training ground and that candidates often get noticed through strong work they publish. He argues that seniority matters little in today's LLM world, and that being at the cutting edge outweighs external social media visibility. He also disagrees with the view that being a middle author on many papers is a negative signal.

AIAnthropic CEO Dario Amodei announced a new companion essay to Machines of Loving Grace, his earlier piece on what powerful AI could achieve if developed well. The post links to the original essay, which he wrote over a year ago.

AIGeoffrey Hinton recommends that every politician watch a conversation about the future of AI before arguing that regulation will hinder innovation. He links to a YouTube video but does not name the participants or the specific points discussed.

AICursor introduced Cursor Blame, a feature that records the "why" behind code changes so teammates can trace code lineage. The company says this context can also help future agents make better decisions, as part of its agent-first rethinking of software engineering primitives.

AIGuillaume Lample, owner of the Mistral account, shared two job postings on Lever, one for a Master's position and one for a PhD position. The post provides only the links and no details about the roles, requirements, or research focus.

AIYi Tay posted that enjoying the work is the secret sauce to AGI, responding to Logan Kilpatrick's remark that having fun is his competitive advantage. The post offers no technical details, figures, or products.

AIAndrej Karpathy's year-in-review lists the LLM paradigm changes he found most notable in 2025. He highlights Reinforcement Learning from Verifiable Rewards (RLVR), which drove most capability gains as labs ran longer RL training, and describes LLM intelligence as jagged, strong in verifiable domains and weak elsewhere. He also covers Cursor-style LLM apps, Claude Code running on the user's computer, vibe coding, and the case for a visual LLM GUI.

AIChatGPT's weekly users in Latin America have nearly tripled year over year. The post credits ChatGPT Go with making it easier for more people in the region to use the service.

AIARC Prize reports that the top Kaggle entry reached 24% on the ARC-AGI-2 private dataset at $0.20 per task, and that all winning solutions and papers are open source. The top verified commercial model, Opus 4.5 (Thinking, 64k), scored 37.6% at $2.20 per task, while a Poetiq refinement on Gemini 3 Pro reached 54% at $30 per task. The author argues that refinement loops are the main driver of 2025 progress, and says ARC-AGI-3 is planned for early 2026.

Why it matters: The post links 2025 competition results to a broader argument about refinement loops, showing how benchmark outcomes are being read as evidence of AI reasoning progress.

AIQuoc Le announced that Google's Gemini team is recruiting for a new Singapore team focused on advanced reasoning and LLM reinforcement learning. The post invites exceptional engineers and researchers to join the team, which is led by Yi Tay and reports into Quoc Le's Mountain View group.

AIIlya Sutskever says scaling current AI methods will keep producing improvements and will not stall. He adds that something important will still be missing.

AIReflection says it will be at NeurIPS in San Diego next week, with its team staffing a booth and a panel on open AI ecosystems. The panel features speakers from Reflection, Meta, AI2, Berkeley's Ray, and SGLang, and will discuss what building open ecosystems requires, lessons from China, and the meaning of transparency and sovereignty.

AIKarpathy argues that verifiability, not specifiability, is the most predictive feature for AI automation, since verifiable tasks can be optimized directly or through reinforcement learning. He says a task is suited to this approach when the environment is resettable, efficient, and rewardable. This explains the jagged frontier of LLM progress, with verifiable domains like math and code advancing rapidly while creative and strategic tasks lag behind.

AIChip Huyen argues that many AI product failures stem from user experience, data quality, and organizational structure rather than the AI itself. She cites a chatbot whose traction improved after adding pre-populated questions and a voice option for users whose hands were busy, and a lead scoring model that was broken because marketing wasn't asking the right questions. She also notes that senior engineers gain the most from AI coding while resisting it more, and recommends building small tools for daily frustrations to solve the "idea crisis."

AIA hiring manager told Chip Huyen that a software engineering candidate who has not experimented with vibe coding is a red flag. Huyen posted the remark on X and asked for readers' thoughts.

AIGeoffrey Hinton recorded a podcast with Jon Stewart, whom he describes as a longtime hero, focused on explaining how AI works. Hinton says Stewart was especially eager to understand the underlying mechanics.