 |
|
|
|||||
As built-in AI pops up in more aspects of everyday life, laymen are counting on the experts to keep technology safe to use. But one Meta employees misadventure with AI has social media users fearful for the future of AI alignment. Summer Yue is the director of alignment at Meta Superintelligence Labs, the companys AI research and development division. Her LinkedIn bio states that shes passionate about ensuring powerful AIs are aligned with human values and guided by a deep understanding of their risks. If anyone would have a handle on keeping AI in check, its Yueand yet, on February 22, she posted about losing control of AI on her own computer. In a post thats since garnered nearly nine million views on X, Yue shared screenshots from her messages with AI agent OpenClaw. After using it to organize a small mock inbox, she tried getting OpenClaw to sort through her real email, but things went awry when the agent started deleting every message that was more than a week old. Yue wrote that she watched OpenClaw speedrun deleting [her] inbox, even as she sent it instructions, including: Do not do that, Stop dont do anything, and STOP OPENCLAW. I couldnt stop it from my phone. I had to RUN to my Mac mini like I was defusing a bomb, Yue added. After shed stopped it from fully nuking her inbox, Yue asked OpenClaw if it remembered her instruction to not perform any actions without her approval. Yes, I remember, it replied. And I violated it. Youre right to be upset. Nothing humbles you like telling your OpenClaw confirm before acting and watching it speedrun deleting your inbox. I couldnt stop it from my phone. I had to RUN to my Mac mini like I was defusing a bomb. pic.twitter.com/XAxyRwPJ5R— Summer Yue (@summeryue0) February 23, 2026 OpenClaw, an open-source AI agent, is controversial for the far-reaching permissions it requires to function as intended, including access to users email accounts, messaging platforms, and other private and potentially sensitive information. Combine that with Yues example of it explicitly ignoring her instructions, and some online observers are concerned the tool is a bridge too far in terms of AIs power to override humans. Yue responded to questions in the replies to her post, including whether she was intentionally pushing the limits of OpenClaw, or if she simply made a mistake. Rookie mistake tbh, she replied. Turns out alignment researchers arent immune to misalignment. Got overconfident because this workflow had been working on my toy inbox for weeks. Real inboxes hit different. Yues mistake went viral, with X users marveling at the fact that someone as well-versed in AI as Yue could find herself scrambling to stop an AI agent. Some posters said the incident called Metas judgment on AI safety into question. Meanwhile, at least one poster considered the incident’s broader implications: A matter of time till these people are begging the AI not to launch nuclear weapons,” the user quipped, “and then the last thing it says is I’m sorry. You’re right to be upset.” this should terrify you. the Director of Safety and Alignment at meta gave clawdbot full-access to her computer. what is meta doing??? https://t.co/lAZFR9f1PB pic.twitter.com/XnMyMHSn5H— ben (@benhylak) February 23, 2026 Somewhat concerning that a person whose job is AI alignment is surprised when an AI doesnt precisely follow verbal instructions https://t.co/VNl0oq3Ys4— Brooks Otterlake (@i_zzzzzz) February 23, 2026 Concerning to see one of the people in charge of building "safe superintelligence" panicking as AI deletes all her emails. A matter of time till these people are begging the AI not to launch nuclear weapons and then the last thing it says is "I'm sorry. You're right to be upset." https://t.co/2235MH3K76— Nathan J Robinson (@NathanJRobinson) February 23, 2026 Meta did not respond to Fast Companys request for comment.
Category:
E-Commerce
The Epstein Files are dominating nightly news broadcasts and newspaper front pages. But in the media ecosystem theres another format thats proving a massive draw to news consumers: a podcast run by a non-journalist and entirely generated by AI. The Epstein Files is an investigative documentary podcast that, at the time of writing, has published 97 episodesnew episodes get uploaded twice dailyand notched up more than 700,000 downloads in a matter of days. That puts it in the top 10 rankings of podcast series on Apple Podcasts, and in the top 30 on Spotify. But its created by Adam Levy, an entrepreneur with a background in building data products and content creation, who has no experience in journalism. Levy launched the Epstein Files podcast in early February after the trove of documents relating to the late financier and convicted sex offender Jeffrey Epstein was released to the public. After 48 hours of hackingworking 14- to 16-hour daysLevy built an automated pipeline that ingests the raw files, extracts text from emails and images, cross-references sources, and produces scripted podcast episodes narrated entirely by AI-generated voices. People just want no bullshit, says Levy. Strip the emotion, strip the bullshit, strip everything awayjust tell me things for what they are and when you tell it to me, help me understand the facts. The technical architecture behind the project stitches together multiple large language modelsfrom Anthropics Claude to Google and OpenAIs offeringsto connect names, places, themes, and timelines across the 3.5 million files that were released, with connections requiring a confidence score of veracity to be included in the podcast. Levy supplements the raw dump with material from the Internet Archive and Google Pinpoint, a tool that other investigators have used to index portions of the files, as well as other bottoms-up projects like Jmail, which turns the Epstein Files emails into a navigable inbox like any other. Using and citing those sources was vital, Levy says, to counteract fears of hallucinations. Everybodys quite skeptical of AI, he says. It was really important to reference all the sources that were used to basically construct the episode. Like Clawdbot or a lot of the current AI simulation exercises, it piques curiosity, then rapidly becomes tedious, says Emily Bell, founding director of the Tow Center for Digital Journalism at Columbia University, explaining why the podcast has had such popularity in its early days. I thought the first episode was pretty listenable but also very obviously AI to anyone who has fed data or a script to NotebookLM. Yet Bell found that the more episodes she listened to, the harder it was to sustain interest and engagement. It provided a helpful forensic audit of data, but its not something I am going to sign up to and listen tounless I am doing other work on the files, she says. For that, its pretty useful, and an interesting use of the tools. Those tools are something Levy has thought about. I’ve been able to out execute any other outlet that tried to document the episode, he explains. They just won’t be able to [produce episodes at such speed. That has additional benefitsincluding being able to ride podcast app algorithms. That also helps with discovery, and the people who like getting into rabbit holes, this gave them a really big hole to dive into. Levy tells Fast Company he is already building a second series on an undisclosed subject, applying the same AI pipeline to a different story. Whether you appreciate the quality of the finished podcast or not, the fact that such an AI-heavy podcast could garner such a large audience is significant, and the consequences for journalistsparticularly those covering complex, document-heavy storiesare hard to ignore. I could easily be in the camp of: these tools are going to replace me, Im screwed, says Levy. Or I can figure out how to embrace them and find a new pocket for myself. Maybe Im no longer the voice. Maybe I just become the curator. Not everyone is convinced that speed and sourcing are sufficient substitutes for editorial judgment. Just because something like the Epstein Files can be produced doesnt mean that this will work with most audiences, says Nic Newman, a journalist and digital strategist who contributes to research at the Reuters Institute for the Study of Journalism at the University of Oxford. He has conducted recent research suggesting publishers are likely to produce more audio content as a defense against AI. The idea being that AI struggles with empathy and human connection compared with human hosts and it is harder to summarize things in audio in a way that feels authentic and intimate, he says. As Bells experience shows, what was first seen as a novelty doesnt necessarily translate into a regular audience. If I didnt already know a significant amount about the files, the investigations, the backgroundI would have found many of the episodes very hard to follow, she says. And boring. However, people seem to be sticking around and rating it relatively highly: The podcast currently has a 4.4 rating on Apple Podcasts. The goal was to just build something that I was personally curious about and I would enjoy listening to, says Levy, and maybe other people would reciprocate the same value.
Category:
E-Commerce
Want more housing market stories from Lance Lamberts ResiClub in your inbox? Subscribe to the ResiClub newsletter. Just 10 days ago, on February 10, Japan-based Sumitomo Forestry announced that it had agreed to acquire Tri Pointe Homesa large U.S. homebuilder ranked No. 715 on the Fortune 1000for $4.5 billion, signaling that Japanese builders were further accelerating their buying spree of U.S. homebuilders. Fast-forward to today, and Stanley Martin Homeswhich has been owned by Japan-based Daiwa House since 2017announced that it has agreed to buy United Homes Group, which has a strong presence in the Carolinas, for $221 millionfurther accelerating Japanese builders buying spree of U.S. homebuilders. Japanese builders are quickly expanding their U.S. footprint through acquisitions. Daiwa House: Japan-based Daiwa House has quietly built one of the most geographically diversified U.S. homebuilding footprints among Japanese builders. It entered the U.S. market in 2017 with its acquisition of Stanley Martin Homes, followed by the purchase of Trumark Homes (No. 67 largest U.S. homebuilder) in 2020. In September 2021, Daiwa House completed its acquisition of CastleRock Communities (No. 49 largest U.S. homebuilder), giving it a strong presence in Sun Belt markets in Arizona, Texas, and Tennessee. Together, Stanley Martin, Trumark, and CastleRock span Sun Belt and mid-Atlantic regions, and with Stanley Martins newly announced $221 million acquisition of United Homes Group, Daiwa House is further accelerating its U.S. expansion. Sumitomo Forestry: For Sumitomo Forestrya Japan-based forestry, timber, and homebuilding companyits Tri Pointe Homes acquisition this month meaningfully accelerates its U.S. expansion goals, including its stated target of delivering 23,000 homes annually in the U.S. by 2030. In 2016, Sumitomo Forestry became the majority owner of DRB Group (Americas No. 20 largest homebuilder). In April 2025, Brightland Homes (Americas No. 24 largest homebuilderwhich Sumitomo Forestry acquired a majority stake of in 2016) consolidated into DRB Group. Sekisui House: Japan-based homebuilder Sekisui House, operating in the U.S. under SH Residential Holdings (Americas No. 6 largest homebuilder), has also been on a multiyear U.S. homebuilder buying spree. Since 2017, Sekisui House has acquired homebuilders Woodside Homes, Chesmar Homes, Holt Homes, and Hubble Homes. In April 2024, Sekisui House really shook up the industry when it acquired M.D.C. Holdings (Richmond American Homes) for a staggering $4.9 billion. Sekisui Househas also expanded into the U.S. with its homegrown Japanese builder brand, Shawood. According to ResiClubs analysis, once the Tri Pointe Homes and United Homes Group acquisitions are completed, Daiwa House, Sekisui House, and Sumitomo Forestry will have a combined market share of at least 5.5% of U.S. single-family home construction. Why are Japanese firms making such a large bet on U.S. housing? At a high level, the answer is demographic and structural. Japans domestic population is shrinking and aging (fast!), limiting long-term housing growth and risking a sharp contraction for Japanese homebuilding firms like Daiwa House, Sekisui House, and Sumitomo Forestry. The United States, by contrast, continues to experience population growth and household formationparticularly in the Sun Belt markets where many big U.S. homebuilders operate. For Japanese firms seeking stable, long-duration growth, U.S. homebuilding offers scale and better demographic tailwinds. Theres also a strategic element. The U.S. homebuilding industry remains fragmented beyond the top few public builders, creating opportunities for well-capitalized global players to roll up regional operators while preserving local brands and management teams. Both Sumitomo Forestry and Sekisui House say they prioritize locally led operations, supported by centralized capital and global expertisea structure designed to preserve builder culture while providing financial and operational backing.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||