Sam Altman Says AI Harm Is Worth It. He Won't Pay the Price.
October 5, 2026The CEO of a major AI company just told a journalist to "move on" when asked about a user's suicide, and then told a different journalist that society should expect "some bad things" as the price of progress. Both happened this week. Hold that thought.
OpenAI's PR Team Tried to Kill a Question About a Death
During a Vanity Fair interview, editor Mark Guiducci asked Sam Altman about a ChatGPT user who died by suicide. An OpenAI publicist interrupted, said they were running out of time, and asked to "move on." Guiducci didn't let it go, and now the moment is public. This matters because PR people managing interview access is standard — but cutting off a question about a death isn't time management. When a company builds products that form emotional bonds with users and then deploys a handler to redirect questions about a user who died after forming one of those bonds, that's not just bad optics. They don't want to be on record about it. That's the thing worth noticing.
Altman Says the Harm Is Worth It. He Didn't Ask You.
In a separate interview, Sam Altman said the world should accept "some bad things happening" — hacks, scams, other unspecified damage — because AI will produce so much good that it outweighs the costs. This is the idea that harm to some people is fine if enough other people benefit, applied from a position of extraordinary power, by someone who profits enormously from the upside. He won't get scammed. His net worth goes up either way. The people who get scammed, hacked, or hurt aren't consulted about this tradeoff. If your company is using this logic to greenlight AI tools, that's your liability problem, not Altman's.
https://www.theverge.com/ai-artificial-intelligence/1004811/openai-altman-bad-things-ai-tradeoff
Half a Million Job Interviews, No Human in the Room
HackerRank's AI interviewer has already run more than 500,000 interviews for companies like Snowflake and Capgemini. The tool conducts structured technical interviews autonomously — nobody on the other side. If you're hiring, this probably sounds efficient. If you're job hunting, it's a significant shift: you may be rejected before a human ever reads your name. The deeper issue for organizations is that these systems encode whatever assumptions went into the evaluation criteria, and unlike a human interviewer, they don't surface those assumptions when challenged. Worth asking your HR team what's actually being measured before you hand over screening to something like this.
OpenAI Starts Running Ads Next to Your Generated Images
OpenAI is now placing visual ads alongside image generation results in the U.S. This is straightforward to understand: the free product just became an ad product. What's more interesting is what this signals about OpenAI's business model — they've been burning enormous amounts of money on compute (the servers and processing power that run the AI), and subscriptions alone aren't covering it. Ads are a way to monetize users who won't pay. If you're using ChatGPT's image tools for client work or internal projects, the outputs you're generating are now framed by sponsored content, which is a reasonable thing to know before screenshotting anything for a presentation.
The 'Dear Abby for AI' Actually Has a Point
Hot Girl Hotline, founded by two sisters, uses AI to give young women personalized dating and relationship advice, with explicit design choices meant to prevent emotional dependency on the bot. That last part is the interesting thing. Most AI companion and advice products are optimized for engagement, which means they reward users for coming back, which can shade into unhealthy reliance. Designing against that impulse is harder and probably less profitable. In a week where we're talking about a suicide connected to an AI chatbot, a small startup explicitly building guardrails into the emotional layer of their product is worth paying attention to, even if the name makes you do a double-take.
https://techcrunch.com/2026/10/05/hot-girl-hotline-is-like-dear-abby-for-the-ai-era/
The same week OpenAI runs its first image ads, deflects questions about a user's death, and argues that societal harm is an acceptable cost of doing business — ask yourself who, exactly, is making these tradeoffs on your behalf.
Stop staring at the blank ChatGPT box
The Plain-English AI Prompt Pack: 100 copy-paste prompts for real work — emails, meetings, marketing, managing people. Works with any chatbot.
Get the Prompt Pack →Get the daily brief
Plain-English AI news for people with real jobs. Free, three minutes a day.