Artificial intelligence is more a part of our lives than ever before. While some might call it hype and compare it to NFTs or 3D TVs, generative AI is causing a sea change in nearly every part of the technology industry. OpenAI’s ChatGPT is still the best-known AI chatbot around, but with Google pushing Gemini, Microsoft building Copilot, and Apple adding its Intelligence to Siri, AI is probably going to be in the spotlight for a very long time. At The Verge, we’re exploring what might be possible with AI — and a lot of the bad stuff AI does, too.
After launching across the US and India, Google is bringing its revamped Finance app to more than 100 countries, including Australia, Brazil, Canada, Indonesia, Japan, and Mexico. Users in these countries can now interact with the app in their local language, as well as access a built-in Gemini chatbot, new charting tools, and an upgraded news feed.
OpenAI is juggling public controversies, strategy shifts, and increasing competition.
Muck Rack collected millions of responses from ChatGPT, Gemini, and other AI platforms to try to measure which news outlets and writers LLMs tend to cite the most. The data, as reported by Press Gazette, is fascinating: niche and little known publications seem to be showing up frequently (along with people like Jim Cramer at CNBC). It’s an AI visibility rat race out there.
The web hosting platform is working with Cloudflare to integrate the company’s AI Crawl Control tool, which lets publishers choose how web crawlers can access their site. Publishers can use the tool to permit or block bots, or ask them to pay.
Cloudflare began blocking AI crawlers by default last year.
As Iran threatens to destroy OpenAI’s Stargate data center in Abu Dhabi, just think of all the extra admin it’s created for the people in charge of construction.
miakizz:
barring anything else, pour one out for some poor construction project manager who suddenly has to add “site being blown up” to the Asana board
Get the day’s best comment and more in my free newsletter, The Verge Daily.
Kozic will focus on infrastructure at the well-funded startup, which Bezos leads with former Google exec Vikram Bajaj, according to the FT. Project Prometheus is focused on using AI to improve manufacturing. Kozic’s defection is the latest in a broader wave of AI talent reshuffling.
The “multiple gigawatts of next-generation TPU capacity” are expected to come online beginning in 2027 to “power our frontier Claude models.” The company also says that its run-rate revenue has surpassed $30 billion.
Google AI Edge Eloquent is a new live AI transcription app that requires no subscription and has no usage limits. When you finish speaking, it will also filter out filler words like “um.” It’s currently only on iOS, but Google plans to bring the app to Android and macOS.
Outage reports have gone up and down on Downdetector today, and my feed occasionally hasn’t loaded. Bluesky says the issue, which it blamed on an upstream service provider, “appears” to have been fixed, but users have taken the problems as a chance to poke fun at unpopular comments from the Bluesky team about vibe coding.
Updates: Added Bluesky’s status updates.
Meta will “eventually” offer open source versions of its new AI models Alexandr Wang is in charge of, but first, the company “wants to keep some pieces proprietary and to ensure they don’t add new levels of safety risk,” Axios reports.
The AI-powered compliance startup is no longer listed on YC’s directory after an anonymous report alleged Delve “fakes compliance” and leaked audit reports, as reported by TechCrunch. Delve responded by claiming a bad actor “maliciously exfiltrated data” as part of a “coordinated, targeted cyberattack.”
A long, and at times funny, report in The New Yorker on Altman’s will to power, people-pleasing, and alleged pattern of deceit, compiled from notes, memos, and more than 100 interviews. Altman’s reputation has given rise to grimmer rumors – hiring sex workers, the sexual pursuit of minors, even involvement in murder – that The New Yorker found no evidence for. Increasingly, the question is not whether computers are intelligent but whether OpenAI’s leadership is.
[The New Yorker]
The head of the networking giant on energy, infrastructure, and why AI is writing Cisco code.
It makes it easy to flood streaming with AI Beyoncé ripoffs.
We’re still grappling with the impact chatbots are having on younger people. But most of the attention is on higher-profile models like ChatGPT, Claude, and good ol’ MechaHitler. But there’s a whole world of role-playing chatbots like Character.ai that have quietly exploded in popularity. According to the New York Times:
He [Quentin] enjoyed harassing the bots with “funny violence,” he said, like running them over with a lawn mower, inflicting harm in an environment with no actual victims. He also created elaborate story lines in which he fought or flirted with his favorite characters. Occasionally, he would indulge in what he called “devious acts” on a platform now called PolyBuzz that offered more sexually explicit chatbots.
It never once told me to walk into a river.
AI-generated ‘Expert Reviews’ weren’t a hit with users or experts.
The contract still needs to be ratified by union members, but it reportedly includes increases to funding for the WGA’s health plan and pension, as well as increases on residuals for streaming. It also bolsters protections against works being used to train AI.
[The Hollywood Reporter]
Murphy Campbell plays public domain ballads, but YouTube accepted the copyright claim anyway.
Human creators want an ‘AI-free’ label, but can’t agree which one.
Meta has paused work with the company, Mercor (which The Verge has profiled), while OpenAI is investigating the security incident, Wired reports.
But apparently when you write a story about something, sometimes your profile photo comes up in the image search results for it, and a major X account doesn’t bother clicking through to figure that out.
Luke Dicken posted about the changes on LinkedIn, as reported by Game Developer, saying that the team’s time “has come to an end.” Before becoming Take-Two’s head of AI, he held the role of senior director of applied AI at Zynga, which is owned by Take-Two, for two years.
The New York Times reports that Elon Musk is demanding that “banks, law firms, auditors and other advisers” working on the SpaceX IPO buy subscriptions to Grok, which is technically now under the SpaceX umbrella.
[The New York Times]
A fascinating profile on litigator Jay Edelson, a longtime tech adversary who’s been filing cases against OpenAI and Google over their LLMs. “Courts are fed up with these companies, and juries are kind of sick of big tech for doing a lot of damage to society,” Edelson says. Sam Altman has called him a “leech tarted up as a freedom fighter,” and Edelson says Altman is “Lex Luthor.”
A proposed class action lawsuit claims Perplexity “effectively planted a bug” on users’ computers by embedding trackers from Meta and Google inside its AI search engine, as reported earlier by Ars Technica. It also alleges that Perplexity’s incognito mode “does nothing” to protect user privacy:
Even paid users who turned on the “Incognito” feature still had their conversations shared with Meta and Google, along with their email addresses and other identifiers that allowed Meta and Google to personally identify them.
Previous versions used a custom license that has been criticized as too restrictive. With Gemma 4, Google is moving to the Apache 2.0 license, which is much more permissive and widely used by developers, including for other Google products like Android. The new model also offers performance improvements, as detailed in the video below.
