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, AI is causing a sea change in nearly every facet of life that technology touches. Bing wants to know you intimately, Bard wants to reduce websites to easy-to-read cards, and ChatGPT has infiltrated nearly every part of our lives. At The Verge, we’re exploring all the good AI is enabling and all the bad it’s bringing along.
Featured stories
If you can’t tell the difference between AGI and RAG, don’t worry! We’re here for you.
On The Vergecast: selfie surveillance, DIY jaywalk detectors, and the artwork of one man with questions about AI.
One thing about having the idea of AI clones attending meetings in Zoom presented to you for the first time in a conversation with the CEO on your podcast is that other people get to react to said idea in a much funnier way, like Angela Collier does here.
How we use the internet is changing fast thanks to the advancement of AI-powered chatbots that can find information and redeliver it as a simple conversation.
A new option spotted in the Chrome 128 beta lets you search with Google Lens by clicking and dragging a box around the area of a website you want more information about. Google will then pull up search results based on the image or text you’ve highlighted — sort of like Circle to Search.
Wired editor-in-chief (and notable Verge alum!) Katie Drummond flags this incredible stack of sourcing and disclosure notes in Semafor’s piece about Perplexity’s new revenue-sharing agreements with publishers. Well done, all around.
Meta’s CEO got a little heated while talking with Nvidia CEO Jensen Huang at the SIGGRAPH conference today in Denver.
The topic turned to Meta’s approach to AI with Llama. Zuckerberg made clear that investing so much in foundational models is strongly influenced by not wanting to relive his history with Apple and the App Store:
“One of my things for the next 10 or 15 years is I just want to make sure we can build the fundamental technology that we’re going to be building social experiences on. Because there have just been too many things that I’ve tried to build and then have just been told, ‘Nah, you can’t really build that,’ by the platform provider that, at some level, I’m just like, ‘Nah, fuck that.’”
The text-to-image generator, which is trained on Getty’s stock pictures, now uses an upgraded version of Nvidia’s Edify AI model, making it faster and more accurate. It also comes with new features to control the “camera settings” used in an AI-generated image, such as depth of field or focal length.
A report from 404 Media reveals how AI companies like Anthropic are bypassing a website’s robots.txt file by deploying new web crawlers with different names. This makes it more difficult for websites to block crawlers, as they constantly need to update their files to include the new bots:
Anthropic’s current and active crawler is called “CLAUDEBOT.” Neither Reuters nor Condé Nast, for example, blocks CLAUDEBOT. This means that these websites—and hundreds of others who have copy pasted old blocker lists—are not actually blocking Anthropic.
Autonomous Cars
Elon Musk is not answering the most important questions about the Tesla robotaxi
Alphabet will invest up to $5 billion in Waymo.
Tandem drifting Toyotas show how AI might help drivers on slippery roads
GM ditches Cruise’s custom-designed driverless car
Mark Gurman reports for Bloomberg that “Apple Intelligence” features will be available for developers to beta test this week.
However, he also says the first ones won’t be released publicly until weeks after Apple’s big September updates for iPhone / iPad / Mac, etc. Rollouts for others from its WWDC showcase, like upgraded Siri, could stretch into 2025.
If AI can learn everything about everything, can it tell me what to watch on Netflix tonight?
Google invented a lot of core AI technology, and now the company’s turning to Demis to get back in front of the AI race for AI breakthroughs.
If you still use X, then you were likely opted-in to have your posts used as training data for the xAI chatbot Grok, Elon Musk’s AI company.
This may violate EU GDPR rules, which require companies to obtain consent before using personal data. The EU privacy watchdog told The Financial Times they were “surprised” and are “seeking clarity” on the issue.
[Financial Times]
With open source driving the cost of AI models down, attention is turning to the products they power.
It’s the latest company to commit to The White House’s voluntary AI agreement, which promotes the safe and responsible development of AI. OpenAI, Amazon, Google, Microsoft, Meta, Adobe, Nvidia, and others already signed onto the agreement last year.
The trend of hallucinations showing up in public AI demos continues. As noted by a couple of reporters already, OpenAI’s demo of its new SearchGPT engine shows results that are mostly either wrong or not helpful.
From The Atlantic’s Matteo Wong:
In a prerecorded demonstration video accompanying the announcement, a mock user types music festivals in boone north carolina in august into the SearchGPT interface. The tool then pulls up a list of festivals that it states are taking place in Boone this August, the first being An Appalachian Summer Festival, which according to the tool is hosting a series of arts events from July 29 to August 16 of this year. Someone in Boone hoping to buy tickets to one of those concerts, however, would run into trouble. In fact, the festival started on June 29 and will have its final concert on July 27. Instead, July 29–August 16 are the dates for which the festival’s box office will be officially closed. (I confirmed these dates with the festival’s box office.)
According to a post from OpenAI CEO Sam Altman, ChatGPT Plus subscribers will get access to its new voice feature next week. The company initially demoed this feature back in May, you know, the one that sounded like Scarlett Johansson’s character from the movie Her? They yoinked that voice option though, sorry guys.
This will be interesting to test out, especially ahead of ChatGPT getting baked into Siri.
Two specialized AI systems from DeepMind, AlphaProof and AlphaGeometry, were able to solve four out of six problems from this year’s International Mathematical Olympiad (IMO). This is a big deal because AI is really bad at math. Have you seen the posts about ChatGPT saying 9.11 is bigger than 9.9?
The IMO represents a “litmus test” that suggests researchers could make bigger future breakthroughs, DeepMind VP David Silver said in a press briefing.
Researcher Alex Hanna, of Distributed AI Research Institute and previously of Google, reflects on what might come next:
After the dust settles and NVIDIA has stopped churning out shovels (e.g. H100s) for the gold rush, what will be left behind? Will data centers go the way of shopping malls? Likely not—they’ll be repurposed for other massive computing projects. But what about those climate pledges?
[Mystery AI Hype Theater 3000: The Newsletter]
Look, I know unsightly stickers on Windows laptops have been with us since long ago when “Intel Inside” was something people actually wanted boast about. But come on, HP.
This is what we’ve pieced together about her views on AI, privacy, antitrust and more.