The name of the company purchased by OpenAI for 6.5 billion dollars in late may, has, for all its hype at the time, been taken down by a court order.
Continue reading “OpenAI takes down Ive partnership Io page due to court order”
Tag: openai
Meta tried to buy Ilya Sutskever’s Safe Superintelligence, now going for execs
They rebuffed Meta’s buyout offer, CNBC writes, claiming information from anonymous sources close to the transaction.
Now, Meta has turned their eyes to recruiting the co-founder of Safe Superintelligence and the previous AI chief at Apple, Daniel Gross, and former GitHub CEO Nat Friedman, who run an AI venture fund together— which Meta would hold a stake in as part of the deal.
Continue reading “Meta tried to buy Ilya Sutskever’s Safe Superintelligence, now going for execs”
Friday news roundup – what you might have missed
YouTube’s owner sits on an archive of over 20 billion videos, now being tapped to train its Veo 3 video model. But creators are alarmed about their content appearing in AI-generated outputs, raising red flags over copyright and IP rights. Google says it only uses a small subset of the repository and has «guardrails» in place:
«we’ve invested in robust protections that allow creators to protect their image and likeness in the AI era,» they say to CNBC.
Midjourney’s new Video model is only $10/month
While other video generation tools come with premium price tags, Midjourney’s new model is pitched as the first truly accessible option: «The first video model for everyone.» It animates either uploaded photos or AI-generated art, comes with a prompt field for control, and outputs 5-second clips. Early reactions: it excels at cartoons and stylized animation.
Read more at The Verge, their launch post on X, blog post, and check out the gallery here.
Biology in future OpenAI models getting so good, they pose a «high risk»
OpenAI warns that some successors to its current o3 models will cross thresholds in biological reasoning that trigger a «high risk» classification—raising concerns about misuse in synthesizing harmful materials. «They won’t be able to create bioweapons per se,» said safety lead Johannes Heidecke, «but extra safeguards will be deployed.»
Read more: A report on Axios and the blog post from OpenAI discussing the issue.
Eutelsat is creating a Starlink competitor
Europe is making a bold move to counter SpaceX dominance. Struggling satellite firm Eutelsat has secured $1.55 billion to expand its OneWeb network –€717M of that from the French state, which now holds a ~30% stake. «We must invest now,» Macron’s office said, «or risk dependence on foreign powers.»
Read more: France24 digs deep, Bloomberg is paywalled.
OpenAI announces $200 million DoD deal, launches government group
The first partnership will be a $200 million Pentagon deal to «identify and prototype» AI across a wide range of functions, from healthcare to acquisitions and cyber defense.
National security «and enterprise»
According to CNBC, quoting the The Department of Defense, the awarded contract will deal with «critical national security challenges in both war fighting and enterprise domains.»
Continue reading “OpenAI announces $200 million DoD deal, launches government group”
Altman sees mass job loss — and plans to counter with more AI
Whole categories of jobs will be wiped out, he writes — but then says AI will bring us so many benefits and new policy options we won’t care much:
— There will be very hard parts like whole classes of jobs going away, but on the other hand the world will be getting so much richer so quickly that we’ll be able to seriously entertain new policy ideas we never could before, he writes.
Continue reading “Altman sees mass job loss — and plans to counter with more AI”
OpenAI’s open weights model now coming «later this summer»
The timeline for the release was originally slated for early summer, writes TechCrunch.
Announced to compete with Llama, DeepSeek
An open weights model basically means you can run the AI on your computer, tinker with it, and modify the «weights,» whitch means the leveling, the priorities and instructions of the model itself.
Continue reading “OpenAI’s open weights model now coming «later this summer»”
OpenAI launches o3-pro, drops o3 API prices by 80%
According to OpenAI’s release notes, it scores about 10% better than o3 in select benchmarks.
It is available for Pro and Team users, replacing the o1-pro model. It should be available to Enterprise and Edu users next week.
o3-pro also has access to tools, and can therefore search the web, analyze files, take visual inputs and use Python, OpenAI says.
Continue reading “OpenAI launches o3-pro, drops o3 API prices by 80%”
ChatGPT had «partial outage» lasting seven hours
The outage was affecting everything from the web service, to iOS/Android/Windows and macOS apps, across GPTs, Deep Research, Operator, Codex, Sora and Image Generation.
Intensifying errors
Downdetector collected reports on the outage since 0900 CET/0300 EST, and it seemed to peak in the early hours today.
Continue reading “ChatGPT had «partial outage» lasting seven hours”
OpenAI hits $10 billion in yearly revenue — well ahead of schedule
According to TechCrunch, the revenue streams from consumer products, 500 million active weekly ChatGPT users, API sales, and 3 million paying business customers.
AI industry is booming
This caps six months of remarkable growth for the AI business. Recently, Anthropic reported record revenue growth, having reached $3 billion in revenue so far this year.
Continue reading “OpenAI hits $10 billion in yearly revenue — well ahead of schedule”
OpenAI sees rise in China-based ChatGPT abuse
Out of the ten campaigns identified in OpenAIs new report «Disrupting Malicious Uses of AI», four were from China.
Supercharging influence ops
Chinese groups have used ChatGPT for mostly adversarial influence operations, writes Reuters, generating social media posts on political tops including on a Taiwanese video game, accusations against a Pakistani activist and content related to the closure of USAID.
Continue reading “OpenAI sees rise in China-based ChatGPT abuse”
OpenAI’s Codex now available to ChatGPT Plus users
Codex is the latest coding agent from OpenAI that runs on a modified o3-model.
Super-coding agent
It can generate several instances of code from your prompts, and even run them in a sandbox to select the best/most efficient version.
OpenAI says it can complete tasks autonomously that would otherwise take hours or days to finish, and they are using it themselves to offload repetitive tasks.
The Plus membership for ChatGPT is $20 a month, and Codex launched as a «research preview» in May for Pro users, who fork out $200 a month.
Update: It appears Codex now also has Internet acccess, which is off by default and comes with a stern warning.
See also: teknotum on the Codex launch, and the announcement thread on X.
Altman warns of “scary times ahead,” but is positive about broad future
They discussed several interesting topics, but Altman said »There will be scary times ahead» as AI gets sharper, without mentioning the recent jobs panic specifically.
Brace for a lot of change
The notion, he said, is for OpenAI to release research previews — or incomplete models — early so that the world can think and prepare for the consequences.
Continue reading “Altman warns of “scary times ahead,” but is positive about broad future”
Jony Ive joins OpenAI on breakthrough hardware project
Citing significant progress already done, Sam Altman says «I think it is the coolest piece of technology that the world will have ever seen,» according to The Washington Post.
Continue reading “Jony Ive joins OpenAI on breakthrough hardware project”
OpenAI debuts Codex, an AI coding agent, further disrupting the software industry
OpenAI’s Agents Research Lead, Josh Tobin tells TechCrunch that the company wants software agents to complete tasks autonomously that would otherwise take hours or days to finish, in anything from one to thirty minutes, and OpenAI is already using the model to offload repetitive tasks.
Continue reading “OpenAI debuts Codex, an AI coding agent, further disrupting the software industry”
ChatGPT 4.1 now available in the app and web
OpenAIs coding model was launched in the API only on April 14th, and reduced costs compared to GPT 4.5.
It also comes with a large context window of 1 million «tokens.» These are compute units that roughly translates to 750,000 words — so this model can read and output very large files or codebases.
It should be available under the model selector menu by clicking on «More models.»
UPDATE: It seems that without announcement of fanfare, free users on ChatGPT.com now get 4.1-mini as the new default model, a step up from the previous 4o-mini.
Despite some reports, GPT-4.1 does not appear to be the default for paid users—ChatGPT Plus still opens with GPT-4o as standard
By popular request, GPT-4.1 will be available directly in ChatGPT starting today.
GPT-4.1 is a specialized model that excels at coding tasks & instruction following. Because it’s faster, it’s a great alternative to OpenAI o3 & o4-mini for everyday coding needs.
— OpenAI (@OpenAI) May 14, 2025
Read more: OpenAI’s model release notes, Ars Technica bemoans model confusion, Teknotum on the 4.1 launch.