hotai.newsAn archive of 20 AI news briefs collected on 2026-06-24. Check the original sources for details and updates.
This content still needs to be cross-checked against the original publication page.
Named Jalapeño, the new processor was designed specifically for the unique needs of OpenAI's inference systems.
Source links can be verified on the original siteRead original ↗The core technical bottleneck of current AI detection lies in the fact that we are using AI to check AI, which leads to the inability to clearly determine whether a piece of text was written by a human author or by AI, and no clear explanation or description can be made of this; this is the most critical part of the technical bottleneck.
Source links can be verified on the original siteRead original ↗Zhou Hongyi unveiled "Tulong Feng" and "Yitian Zhen" at the ISC 2026 conference, aimed at countering the cybersecurity dimensional strike brought by the US AI model Mythos. He stated that AI is increasing vulnerability discovery speed by a hundredfold and costs are plummeting, and traditional defense logic faces failure. 360 has built a Chinese version of Mythos through the agent route, and has cumulatively discovered over 3000 vulnerabilities. #AI Security#
Source links can be verified on the original siteRead original ↗OpenAI has revealed its first AI-focused chip made in partnership with Broadcom. Jalapeño is an ASIC designed to power inference for current and future AI models behind apps like ChatGPT.
Source links can be verified on the original siteRead original ↗OpenAI's AI coding tool Codex CLI has a serious flaw that continuously writes massive diagnostic logs to SSD, reaching approximately 640TB per year, causing rapid depletion of solid-state drive lifespan. OpenAI has acknowledged the issue and is urgently fixing it. #AI Coding##SSD Lifespan#
Source links can be verified on the original siteRead original ↗HSBC commissioned Ipsos to survey about 10,000 affluent and high-net-worth individuals. The results showed that 62% of respondents still regard professionals as the main source of investment ideas, and only 12% consider AI the most important factor affecting investment decisions.
Source links can be verified on the original siteRead original ↗The LinkedIn cofounder and investor in both Anthropic and OpenAI offers his most pointed public assessment yet of Elon Musk’s AI ambitions.
Source links can be verified on the original siteRead original ↗The current environment is abnormal. After completing verification, you can continue to access. Go to verification
Source links can be verified on the original siteRead original ↗OpenAI is adding custom hardware to its tech stack. The "Jalapeño" chip, developed with Broadcom, is tailored for large language model inference and is set to run at scale by late 2026.
Source links can be verified on the original siteRead original ↗Modular's AI-native software platform runs AI models on various XPUs with industry-leading performance, without needing to rewrite code for each architecture.
Source links can be verified on the original siteRead original ↗Jalapeño was co-designed by OpenAI engineers and Broadcom, mainly for AI inference. Broadcom CEO Hock Tan said in an interview with Reuters that Jalapeño's performance is comparable to Nvidia's Blackwell chips and Google's TPU.
Source links can be verified on the original siteRead original ↗OpenAI deployment chief Arnaud Fournier explains in an interview how DeployCo wants to embed AI deep inside large corporations using its own engineers. He talks about explosive Codex growth, the feedback loop from customers back into model development, and why he thinks the price of AI intelligence has dropped sharply.
Source links can be verified on the original siteRead original ↗The current environment is abnormal. After completing verification, you can continue to access. Go to verification
Source links can be verified on the original siteRead original ↗Language models may write cleaner prose than most humans, but ask one for 100 arguments on a topic and they'll all cluster together. Human reasoning is far more diverse, says Pangram CEO Max Spero, and that's what might give AI away.
Source links can be verified on the original siteRead original ↗Stanford University is located in the core area of Silicon Valley and holds a special status in the US technology industry. It is not only regarded as an innovation hotbed, but also surrounded by several of the world's most influential tech giants, and is home to many AI pioneer companies. Stanford students also typically have an advantage when entering the job market.
Source links can be verified on the original siteRead original ↗Samsung Electronics is expanding employee access to ChatGPT Enterprise and Codex, giving staff wider use of AI tools for technical and non-technical work.
Source links can be verified on the original siteRead original ↗Abstract page for arXiv paper 2606.24597: Qwen-AgentWorld: Language World Models for General Agents
Source links can be verified on the original siteRead original ↗Anthropiclauncheda beta version of its Claude Tag feature for Enterprise and Team tiers, shifting its chat model into shared Slack channels. Moving away from traditional isolated chat boxes, users pull the artificial intelligence model into active group threads by typing @Claude.
Source links can be verified on the original siteRead original ↗Claude Tag lets teams bring Anthropic's AI into Slack by tagging @Claude in any channel and assigning it tasks. Internally, the tool already generates 65 percent of the code on Anthropic's product team, the company says.
Source links can be verified on the original siteRead original ↗Meituan Vice President and Dianping Business Unit General Manager Tao Xuexuan said in a media exchange that compared with larger internet products, Dianping's reviews are more like a community bulletin board.
Source links can be verified on the original siteRead original ↗