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.
When making an interactive experience with Project Genie, you can now pick a place in the US and Genie can use Street View data from that place for grounding. You can also prompt a style to change what that place looks like in your interactive AI-generated world.
It’s still expensive, though: Google AI Ultra plan, which used to cost $249.99 per month, now starts at $100 per month. There’s also a $200 per month option that includes access to Google’s Project Genie world model.
The two options for Ultra matches OpenAI’s approach for its Pro tier, which also now comes in $100 and $200 per month tiers.
The Federal Trade Commission launched a new website for consumers to report alleged failures to comply with the Take It Down Act, which went into full force today. Critics fear the law will be used to censor online speech, but the website also has resources for domestic violence survivors.
Karpathy, who had also been on the founding team of OpenAI, says he will be working on R&D at Anthropic. Previously, he had been working on “new kind of school that is AI native,” and he says he’s still “deeply passionate about education” and plans to go back to it “in time.”
After launching the Verified by Spotify badge for musicians a few weeks ago, the service now has a sticker for podcasters who actually exist in real life, as AI-generated podslop floods RSS feeds.
It’s also “reaffirming” its unauthorized impersonation policy, saying it will remove podcasts that copy someone else’s likeness without permission, whether through AI cloning or any other method.
The Take It Down Act is in full force, but it could be a gift to government censors — not victims of image-based sexual abuse.
That sparkle icon is showing up everywhere these days.
After launching for iPhone last year, Firefox users on Android can now shake their device on any webpage under 5,000 words to get an AI-generated summary of its contents. You can disable the feature entirely, or select “Summarize Page” under “More” in the three-dot menu if you prefer tapping.
At Bloomberg, Mark Gurman has more AI-related rumors ahead of WWDC, saying that, along with a reworked version of Siri, Apple plans to build Grammarly-like grammar checking and suggestions into the next iPhone and iPad updates (hopefully, without using our AI slopplegangers for an “expert review”).
Other changes include a Shortcuts upgrade that builds automations based on whatever requests you describe, and an AI wallpaper generator similar to what Samsung and Google already offer.
Public opinion of the AI industry is already sinking. A parade of untrustworthy executives makes it look worse.
Sen. Adam Schiff (D-CA) announced the Energy Cost Fairness and Reliability Act, which would put new requirements on “energy-intensive facilities,” in an effort to lower the strain on the energy grid. It doesn’t yet have co-sponsors, but hits on an issue that’s become central to many communities and elections.
That was quick (about two hours).
I assume because he was recording, since the marshal said, “Give me your phone.” There have been several incidents of people attempting to record or take pictures throughout the trial — but I honestly am not sure why you’d record today of all days.
Unfortunately he does not have a lot of details YGR is asking for. He doesn’t know how many investments Musk has made (11 to date, according to Pitchbook), or how many were successful. He’s getting some pretty tough questioning from YGR in the direct exam. Among the things she’d asked, he didn’t know how many startups fail in Silicon Valley, or the success rate for assorted VC firms.
Last week we saw clips of a Cybertruck being rescued on a California beach by a minivan. It’s a funny situation that begs for immediate coverage — like this, from The Drive. But both The Drive and the cited YouTube video hedged that it could be fake. So we also became skeptical and looked for potential AI shenanigans, after examining the license plate on the Cybertruck in the video. I asked the California DMV for comment, and it just got back to us confirming that “the vehicle plate and registration are valid for the plate number you provided.” So, at least it seems like it’s a real car.
AI promises to make lives easier, but for those in the business of truth, it’s making verification more expensive.
During an interview for his new animated film Tangles, at Cannes, Rogen minced no words about his feelings on AI, calling it “stupid dog shit.” But perhaps more pointedly said that, if you’re using AI to write, just “go do something else,” because you’re not a writer.
There are leaderboards for burning AI tokens, but now there are literal receipts. Sascha Corti, an engineer at Microsoft, just shared the implementation — from code designed by Chris Hutchinson (of course, with the help of Claude). Is it cute? Yes. Is it cursed? Also yes.
Replit CEO Amjad Masad announced on Friday that the app got its first iOS update in four months. In March, Apple reportedly blocked Replit and other vibe coding apps from publishing App Store updates unless they made changes, potentially including moving generated app previews to web browsers.
