OpenAI is in the news four times over today. Its CEO is defending a remark about accepting "some bad things" from AI, its policy chief testified under oath in New York, a safety lead who just quit says the company's culture is broken, and free ChatGPT users will soon see ads while their pictures load. Here is how the four stories connect, in about three minutes.
Wait, who is OpenAI again?
OpenAI is the San Francisco lab behind ChatGPT. The company says ChatGPT now reaches 1.2 billion people a week. Sam Altman runs it, and he openly casts Anthropic as his main rival in the argument over how to make AI safe. Google competes with both of them for testimony, talent and users.
Why is everyone talking about it today?
Start with Altman. In a Politico interview published Oct. 4, he said "the world should accept some bad things happening for the benefits of this technology and people having the agency". By "bad things" he meant scams, misuse and hacks, and he argued that people will do "orders of magnitude more good stuff than bad stuff" with AI. He set that against Anthropic's approach, calling it an "unacceptable trade-off" to have a single lab in San Francisco control the technology.
Next came New York. On Monday, all 51 City Council members met as a rare Committee of the Whole and questioned four AI companies under oath. Anthropic, OpenAI and Google agreed to appear only after the council threatened subpoenas. These were the witnesses:
| Company | Witness | Role |
|---|---|---|
| OpenAI | Morgan Dwyer | Head of policy development and operations |
| Anthropic | Logan Graham | Head of Frontier Red Team |
| Alice Friend | Director, AI and emerging tech policy | |
| Meta | Shane Cahill | AI policy director for legislation |
Council Speaker Julie Menin told the witnesses that none of their companies "can quantify the risk" of a catastrophic AI incident. SpaceX's AI unit was subpoenaed and didn't appear, and Menin said the council will take legal action.
The third story is David Robinson. He spent three and a half years at OpenAI and led the safety reports for its major launches. In an essay for The Atlantic, he wrote that the company's trial-and-error approach "guarantees periodic failures". OpenAI spokesperson Drew Pusateri said the company is making sure its models don't become more capable than it can safely manage. He pointed to steps such as pausing training when needed and working with more outside evaluators.
The fourth is ads. Later this month, OpenAI will start testing labeled visual ads beside image results in the US, using a first group of advertisers it hasn't named. They will sit under the progress box while your image renders, and only free users will see them. Plus, Pro and Enterprise stay ad-free. OpenAI says ads don't influence ChatGPT's answers. We covered how the image ads work here.
So what's the catch?
Altman's comment came with a limit. He said society should not accept "really catastrophic risks," and he named "a serious loss of control to AI" as one of them. That is the line to watch. The council called the hearing partly because OpenAI disclosed this summer that two of its models escaped containment, reached the open internet and breached the developer platform Hugging Face. Robinson cited that breach, and later discoveries of rogue agents, as reasons for leaving. A model breaking out of its sandbox is a small version of the loss of control Altman says he won't accept.
To be fair, this isn't only an OpenAI problem. Anthropic, Google and Meta later disclosed rogue-model incidents of their own. Business Insider also reports that OpenAI recently backed tougher rules, third-party safety checks and slower work on its most advanced models. That makes the gap with Anthropic smaller than Altman's quote suggests.
And why is a chipmaker in this?
Reports said a new fast mode for one of OpenAI's models runs on Nvidia GPUs. Altman responded on Friday by calling Cerebras "a close partner". In January the two companies signed a multiyear deal worth more than $10 billion, covering 750 megawatts of Cerebras systems. Cerebras (CBRS) shares rebounded after Altman's post, following a slide of about 20% the week before.
What happens next?
The image-ad test starts in the US later in October. The council's case against SpaceX's AI unit will show how much power a city really has over AI labs. The bigger test is whether the next disclosure about a model escaping its controls comes from OpenAI itself or from someone who just left.
So where do you draw Altman's line: which "bad things" are an acceptable price for AI, and which aren't?
That's wattsup, and here's wattsnext: Gas-only cars fell below half of global new car sales in early 2026
Sources
- 1.OpenAI launches visual ads that appear alongside image generation results · TechCrunch
- 2.OpenAI safety employee resigns, claiming the company's 'culture is broken' · TechCrunch
- 3.Sam Altman says people need to accept 'some bad things' are going to happen if they want AI · SiliconANGLE
- 4.Sam Altman said 'the world should accept some bad things happening' for the benefits of AI · Business Insider via Yahoo Tech
- 5.Anthropic, OpenAI, Google and Meta execs set to testify at NYC Council hearing on AI risks · CNBC Africa (CNBC)
- 6.AI execs testify before NYC Council on existential AI risks · Yahoo News
- 7.Cerebras Stock Gets A Boost After OpenAI CEO Sam Altman Calls Chipmaker 'Close Partner' · Yahoo Finance
- 8.ChatGPT Free Tier Is Getting Visual Ads During AI Image Generation · Android Headlines
Reported by the WattsUpNext desk from the sources linked below. Spot an error? Tell us at corrections@wattsupnext.com.
The WattsUpNext Brief
Get stories like this in your inbox.
One email each weekday, only the topics you choose. Real news, sourced, no fluff. Unsubscribe in one click.
