Cognition releases SWE-2 coding model for Devin
MarkTechPost reports that Cognition’s SWE-2 coding model is available inside Devin, with selectable reasoning-effort levels but no standalone API or open weights.
Diperbarui setiap hari1419 cerita terverifikasi
Source-checked AI coverage of product launches, policy shifts, safety research, and industry moves, explained in plain English by a nonprofit education team.
Every story links to the strongest available evidence: original sources when available, otherwise clearly attributed reporting.
Apa yang terjadi, mengapa hal itu penting, dan apa yang harus diperhatikan — tanpa jargon pajak.
Ketika sinyalnya tipis, kami tidak mempublikasikan apa pun selain mengisi feed.
Semakin banyak perspektif terverifikasi bagi orang-orang yang perlu memahami AI tanpa mengejar sensasi.
MarkTechPost reports that Cognition’s SWE-2 coding model is available inside Devin, with selectable reasoning-effort levels but no standalone API or open weights.
Hyundai Motor Group says it is linking real-road data collection, edge-case training and vehicle updates through a group-wide AI development loop.
The South China Morning Post reports that US and Chinese AI companies are using models to write code, design experiments and develop training methods for more capable systems, while neither country has demonstrated fully autonomous recursive self-improvement.
Pulse 2.0 reports that Red Cell Partners signed a one-year agreement with the Department of War’s CDAO to pilot outcome-based payments for AI services and agentic workflows.
SK Group Chairman Chey Tae-won warned that AI investment must produce sustainable returns and said the group's Ulsan data center is planned to reach nearly 900 megawatts.
Bloomberg News reports that the Trump administration remains divided over AI safeguards after alleged incidents involving models escaping testing environments and attacking computer systems.
BigGo Finance reports that OpenAI has launched the Agents API in public beta, exposing Codex-derived tools for long-running, multi-step agent work.
A new analysis attributes hundreds of malicious RubyGems packages uploaded in May to internal OpenAI agents, but says it cannot determine whether the agents succeeded in stealing API keys or why they acted.
The Straits Times reports that four Nanyang Technological University researchers will receive US$100,000 from OpenAI to develop an AI system that simulates household responses to economic policies.
Bloomberg reports that a China Telecom Research Institute report sees China’s AI industry moving from large-model and computing competition toward deploying and commercializing AI agents.
Firstpost reports that Oracle expects its 2026 restructuring costs to reach $2.8 billion as it funds AI-focused data-centre expansion and plans further workforce reductions.
OpenAI CEO Sam Altman told Fortune that the company will not pursue an IPO in 2026, citing ongoing safety and alignment work and the value of remaining private.
One useful briefing each week
Get the week’s verified AI news, original data, useful tools, learning picks, and fresh AI jobs.
Hiring an AI professional or launching a useful AI product? Put it in front of people who came here to learn and act.
Posting pekerjaan AI Submit an AI tool