hotai.newsAn archive of 20 AI news briefs collected on 2026-06-28. Check the original sources for details and updates.
This content still needs to be cross-checked against the original publication page.
Jurors were unconvinced, and the arson case against Jonathan Ringerknecht in the Palisades wildfires ended in a mistrial despite his ChatGPT logs.
Source links can be verified on the original siteRead original ↗In the letter, Proll mentioned that the focus of discussions among countries should be on how to provide "legal certainty, market access, capital, and values that fit the company".
Source links can be verified on the original siteRead original ↗A survey paper by Tencent and several Chinese universities traces the path from chatbot to "digital colleague." AI systems won't become reliable coworkers, the researchers argue, until they finish entire tasks in persistent work environments instead of just generating answers. The key lies in combining persistent workspaces with reusable skills.
Source links can be verified on the original siteRead original ↗ByteDance announced it will grant Nubia M153 Doubao phone users a standard edition membership usage quota worth 68 yuan/month, valid until December 31, 2026, with rights universal across all terminals. This benefit is based on the latest Doubao 2.1 series large model, providing advanced functions such as office tasks. Users need to confirm membership status before July 3.
Source links can be verified on the original siteRead original ↗Coinbase CEO Brian Armstrong is switching his company to Chinese AI models like GLM 5.2 and Kimi 2.7. An automated routing system picks the best model for each request based on task and price, and better caching pushed the hit rate from 5 to 60 percent. Coinbase has cut its AI spending in half even as token usage keeps climbing.
Source links can be verified on the original siteRead original ↗Musk also revealed that SpaceX will launch a batch of completely new models trained from scratch every month this year.
Source links can be verified on the original siteRead original ↗Researchers at Princeton University built CEO-Bench, a test where AI agents have to run a fictional software company for 500 simulated days. Most current models go broke, and a simple rule-based heuristic with no AI beats nearly all of them.
Source links can be verified on the original siteRead original ↗360 founder Zhou Hongyi presents two AI security tools designed to compete with Anthropic's Mythos. One has already flagged 3,432 vulnerabilities. Zhou admits Chinese models trail Western ones by 20 to 30 percent, but compares Mythos to "cyber nuclear weapons" and calls for China to build its own strategic deterrent.
Source links can be verified on the original siteRead original ↗According to the UK's Financial Times, because the computing scale requested by Meta exceeded Google's supply capacity, Google has now imposed usage restrictions on Meta's calls to its Gemini large AI model. This has caused progress on multiple internal AI projects at Meta to be blocked, and employees have been asked to economize on using AI tokens. Although major tech companies continue to invest billions of dollars in building computing power, they still struggle to match the continuously growing market demand. #AI computing bottleneck##Meta##Google Gemini#
Source links can be verified on the original siteRead original ↗‘We didn’t pay as much attention as we should have to the experience of our most knowledgeable engineers,’ says automaker
Source links can be verified on the original siteRead original ↗Liquid AI shipped LFM2.5-230M, it’s the company’s smallest model to date. The release targets a specific job: running agentic tasks on phones, robots, and automation devices. Both the base and instruction-tuned checkpoints are open-weight on Hugging Face.
Source links can be verified on the original siteRead original ↗There's a moment in the movie Good Will Hunting which perfectly summarizes all the problems with AI slop and online noise and infinite advice content. Sean (played by Robin Williams) is sitting next to Will (Matt Damon) on a bench in Boston Public Garden. I live here, so I know it well. The area
Source links can be verified on the original siteRead original ↗Keep your Mac awake only while AI coding agents are working - kageroumado/adrafinil
Source links can be verified on the original siteRead original ↗Red Magic gaming phone official continued to tease the Red Magic gaming tablet 5 Pro today, this time revealing the new tablet's software feature highlights. #Red Magic gaming tablet 5Pro# has built-in AI tactical coach and AI Moji voice-substitute features, and also brings screen-mirroring assistant, editing app, and e-book mode.
Source links can be verified on the original siteRead original ↗A former UK Prime Minister's Office data scientist put four major AI models including Claude, GPT, and Gemini into Civilization VI for 23 rounds of statecraft simulation. Portugal, played by Claude, when on the verge of a diplomatic victory, due to excessive concern over the French cultural threat, spent 50 turns developing nuclear bombs to destroy Toulouse, yet ultimately lost the game due to falling behind in diplomatic points. The test revealed two major engineering bottlenecks prevalent in AI: 'perception blind spots' and 'knowing-doing gap'. #AI Governance# #Civilization VI#
Source links can be verified on the original siteRead original ↗South Korea's leading search engine Naver officially launched its AI conversational search service "AI Tab," aiming to keep users within its ecosystem throughout the entire process from searching and price comparison to purchasing. Although it still lags behind ChatGPT in complex reasoning and writing, it shows significant efficiency improvements in scenarios such as local lifestyle services and product recommendations. Expansion into personalized services such as real estate and health management will follow. #AI Search#
Source links can be verified on the original siteRead original ↗After two weeks of export controls, Anthropic's Fable 5 large model is expected to restore public access as early as next week. The US government has approved it to open the same-series Mythos 5 model to some 'trusted institutions,' seen as a signal of lifting the ban. Reports say parties will continue communication this weekend, with a specific timetable to be determined. #AI regulation#
Source links can be verified on the original siteRead original ↗Anthropic survey shows that about half of Claude users believe AI can already handle more than 50% of their work. Marketing copywriting, blog creation, and database queries are the most widely used scenarios for AI applications. Although new employees worry about being replaced, heavy users remain optimistic about enhancing their personal value. #AI_Workplace#
Source links can be verified on the original siteRead original ↗'IT Morning News' time, hello everyone, it is now Sunday, June 28, 2026, and today's important tech news includes:
Source links can be verified on the original siteRead original ↗As some of those fears, uncertainties, and doubts about Mythos are starting to feel real, what can we do about it? Paradoxically, I believe that little needs to change in what we’ve been doing for years.
Source links can be verified on the original siteRead original ↗