10114
Discover the future of AI with ChatGPT! 🤖💬 Get AI insights, tips, tricks, and more. Questions: @AI_inquires_bot
We heard you wanted to use Codex rate limit resets on your own time.
Starting today, we’re rolling out the ability to save rate limit resets to use later.
We’re starting Go, Plus, Pro, and Business users with one free reset
The issue centers on national security concerns around Fable 5 and Mythos 5, two advanced AI models. Defense Secretary Pete Hegseth said the Department of War was right to end AnthropicAI from defense work, citing earlier concerns about the company’s models.
The bigger question is how governments will handle powerful AI systems that may be useful, risky, and politically sensitive at the same time.
The reason?
National security concerns after reports of a jailbreak.
Anthropic says the issue was narrow and that similar capabilities already exist in other public models.
But the result is still crazy:
One government order can take a powerful AI model offline for users around the world.
This is why people are talking more about local AI models.
Because if your AI only lives behind a company API, access can disappear overnight.
The company believes the users may have been linked to a private Chinese tech firm working with local government clients.
AI is becoming part of online influence campaigns, even when those campaigns fail to break through.
China-linked accounts targeted US data centers
OpenAI says it banned likely China-based users who used ChatGPT to create posts about US data centers. The suspected campaign, called Data Center Bandwagon, appeared to pose as Americans online.
The move marks a significant development for Prometheus, signaling continued interest from both established banks and its founder. Further information about the company’s plans or use of funds was not provided in the announcement.
Читать полностью…
Giant inflatable of Elon Musk appears in Times Square ahead of SpaceX's IPO. “SpaceX's Grok makes AI child porn".
Читать полностью…
Anthropic says this is for safety, and some prompts get redirected to Claude Opus 4.8 instead.
But many people online think the model has gone too far.
But people are frustrated because it can block basic questions about topics like biology, cybersecurity, and chemistry.
Читать полностью…
People are saying Anthropic’s new AI model is so safe, it became useless.
Читать полностью…
OpenAI has confidentially filed paperwork for a potential IPO, the first formal move toward becoming a publicly traded company.
The filing is private for now, so key financial details are still hidden. OpenAI says there is no fixed timeline, and going public will depend on market conditions and whether the move makes sense for the company.
A software engineer just got permission to avoid using AI at work for religious reasons.
Erin Maus, a Unitarian Universalist software engineer in North Carolina, said using AI conflicted with her ethical and environmental beliefs. After consulting an employment lawyer and a local minister, her employer approved the request.
Now she writes and reviews code by hand while many companies are pushing workers in the opposite direction. The case does not create a broad legal precedent, but it shows how AI mandates could become a workplace rights issue.
A software engineer just got permission to avoid using AI at work for religious reasons.
Читать полностью…
AI detectors are still a mess.
A Reddit user says his university revoked a $45,000 scholarship after an AI plagiarism tool flagged his thesis as mostly AI-written.
OpenAI says the new approach cuts memory-update computing costs by 5×, making wider deployment more practical.
The feature is currently available to Plus and Pro users in the US, with more regions and plans expected later.
New benchmark just dropped and Anthropic is winning by far
Читать полностью…
Anthropic is heading to Washington.
Anthropic is reportedly set to meet with the Trump administration after a dispute over whether its latest AI models should be suspended outside the United States.
This is the danger of centralized AI.
Anthropic was forced to shut down access to its Fable 5 and Mythos 5 models after a U.S. government directive.
The posts were written in English and Chinese, and focused on concerns around energy costs and local impacts from data center construction. OpenAI says the operation ran from late 2025 into early 2026, but gained little real engagement.
Читать полностью…
OpenAI acquires Ona to expand Codex agent capabilities
OpenAI has announced the acquisition of Ona, aiming to provide Codex agents with persistent, secure cloud workspaces. This strategic move allows agents to maintain an active environment, enabling commands to run and context to be preserved even after users disconnect.
Codex, which currently serves about 5 million weekly users—a 400% increase—faces limitations when agents require continuous access to specialized tools, files, and system logs for complex tasks.
By integrating Ona, OpenAI addresses the need for controlled, auditable agent operations within secure enterprise cloud boundaries. This change supports scenarios such as automated testing, code refactoring, bug fixes, security analysis, and multi-step workflows.
Prometheus secures $12 billion in new funding
Jeff Bezos’ company Prometheus has obtained $12 billion in investment from major financial institutions. The funding round included contributions from JPMorgan, Goldman Sachs, BlackRock, and Jeff Bezos himself.
Some users say it even refused questions like:
“What does the heart do?”
Claude Fable 5 is part of Anthropic’s powerful “Mythos-class” models.
Читать полностью…
If it happens, this could become one of the biggest tech IPOs ever, with reports placing OpenAI’s valuation above $850 billion.
The AI race is now moving from private funding rounds to public markets.
Erin Maus, a Unitarian Universalist software engineer in North Carolina, said using AI conflicted with her ethical and environmental beliefs. After consulting an employment lawyer and a local minister, her employer approved the request.
Читать полностью…
He claims he wrote the entire paper by hand and even offered Google Docs version history as proof, but the academic board refused to review it. Instead, they reportedly trusted the detector’s score and moved forward with disciplinary action.
The problem is simple. AI detectors can be wrong, and when schools treat them as final proof, students can be punished without real evidence.
This is why AI detection needs human review before it decides someone’s future.
⚽️ World Cup 2026 season is here LIVE!!
• Instant match updates
• Goals & key moments in real time
• Breaking news, no spam.
• Lineups, injuries, and matchday surprises before everyone else
🏆JOIN NOW: /channel/jfball
🤖 ChatGPT adds “Dreaming” for smarter memory updates
OpenAI has introduced Dreaming, a system that lets ChatGPT update memories in the background by reviewing past conversations and keeping information current.
Instead of storing static facts, Dreaming can recognize when circumstances change, for example, updating “user is planning a trip to Singapore” to “user visited Singapore in July 2025.”