Jack Dorsey's Block announced an AI model called Dessa that's trained "from scratch on nothing but human conversation," and to show off the model, Block shared chats it has had with humans that can occasionally be a little awkward and uncanny.
Take this one, for example, shared by Dessa lead Ragavan Thurairatnam. The conversation between the model and the speaker does have a mostly natural back and forth (though for some reason, the model initially mistakes the speaker for somebody named "Ronaldo").
We made the first AI model you can have a realistic conversation with that was trained entirely on speech!
Hosted by TechCrunch in partnership with Fidelity Private Shares, the full-day event goes beyond fundraising and AI to explore the realities of finding customers, building teams, sharpening a pitch, developing founder leadership, and scaling companies both inside and outside traditional startup hubs.
While we were swooning over “WandaVision” back in 2021, we couldn’t have known Marvel was softening us up for the AI revolution. Well, some of us probably suspected something was afoot since, like Agatha Harkness, it was right there all along. Algorithms were already predicting our tastes, needs and desires and serving us options before we asked for them. Virtual assistants steered our daily lives and still do. But Alexa and Siri don’t have faces, bodies or plausible imitations of human emotions or empathy — not on the level that Paul Bettany’s synthezoid Vision provided for a grieving Wanda Maximoff (Elizabeth Olsen).
One five-year time jump later, “VisionQuest” drops in on the android’s story where “WandaVision” left off. The Vision who lived, laughed and loved beside Wanda was an illusion, a delusion, really, brought to life by witchcraft. This one is very real: a kind of Vision 2.0, rebuilt by the intelligence agency tasked with saving the New Jersey town Wanda took hostage to shore up her classic family sitcom simulation. It – OK, he – has no direct connection to the life his predecessor shared with Wanda or the children they magicked into existence.
But Vision is slowly learning what it means to feel, which makes him the AI romantic companion of everyone’s dreams.
Honestly, that’s the show’s main appeal. Marvel and “VisionQuest” creator Terry Matalas seem to know it. Planting quippy chatter wherein Vision deduces that a young woman finds him more attractive than his sidekick Tommy (Ruaridh Mollica) acknowledges Bettany’s rizz is shouldering the weight of this thing. But it also gives tacit permission to lust after the robot he’s playing. After all, the first Vision, or Wanda’s idea of him, tended to her every need and gave her everything she wanted, including children. He also blessed us with one of the most heartbreaking lines of this decade: “What is grief if not love persevering?” A beautiful notion. Also, it kinda made some people want to hit that.
The new Vision returns to a lonely, disconnected world where AI companions are an increasing part of our lives, if not entirely normalized. Falling for a bespoke arrangement of code remains a novel concept for most people, but it’s also an accessible alternative to isolation.
Lacking sorcery or some other way to conjure up the ideal lover, we’re left with engineering. Reams of magazine profiles spotlight everyday people choosing to share their lives with an AI boyfriend, girlfriend or spouse. You may be acquainted with someone who does this without even knowing it, because it’s still viewed as strange or even alarming.
The “VisionQuest” hype is real, too. Although the ever-expanding Marvelverse collapsed under market saturation, Bettany’s romantasy hero is one of the few characters worth getting excited about. He was part of the TV show that started it all – a pretty great one at that – so his reappearance after years of diminishing returns has sparked excitement and hope in equal measure.
That also means that it doesn’t really matter that the “VisionQuest” narrative favors mechanical efficiency over emotional resonance now we’re four episodes in.
As with all Marvel series, “VisionQuest” hasn’t yet landed on its deeper meaning. It took two episodes to bring Tommy and Vision together, and this week’s third and fourth introduced T’Nia Miller’s Jocasta and her connection to the hero/villain synth we know and love.
And regardless of the metaphorical or philosophical cloak it may eventually wrap itself in, the show’s main function is to bridge one chapter of this franchise to the next. Remember, we’re less than two months out from the premiere of “Avengers: Doomsday.”
Translation: This Vision has a purpose beyond figuring out his role in this world, something he put off by hiding in a pub under the alias of Sherwood Ellis and posing as human. Since the establishment was owned by one of the sweetest ladies in the multiverse, who’s killed in a shootout orchestrated by a mercenary named Paladin (Todd Stashwick), her death prods Vision to do a little soul-searching.
But wait — he has no soul, just wires, processors, fluid and whatnot, plus a set of AI personas guiding his interactions with the human world. Rather, most of them are. The original Vision also absorbed the persona of the genocidal superintelligence Ultron, played by James Spader here and in “Avengers: Age of Ultron.” That Vision kept Ultron’s code intact for reasons beyond sentimentality; he/it is effectively Vision’s father, created when Bruce Banner and Tony Stark merged the code of Tony’s noble assistant, J.A.R.V.I.S., with Ultron and the Mind Stone. It makes about as much sense as Tommy, a human, realizing he is somehow one of Vision’s sons due to a confluence of magic and circumstance. (The other, Billy, made his debut in “Agatha All Along.”)
All this is to say that the by-the-numbers circuitry of Tommy and Vision’s cross-country escapade, which thrusts them into a murder case involving a killer dubbed the Butcherbird (Miller), is merely a skin over the ones and zeros of artificial desire.
Take Spader’s role in this show, for example. His Ultron avatar sports the long beard of a fisherman shipwrecked by a hard life. But there are moms and grandmas watching who still blush at memories of Spader’s work in “Secretary.” (If you know, you know.)
Want more from culture than just the latest trend? The Swell highlights art made to last. Sign up here
So, in the denouement of the fourth episode, when Vision lets Ultron take over his handsome, hella tight chassis to get out of a deadly situation, you can’t help but look at that twist with both horror and straight-up horniness. I mean, forget Reese’s cups – for some, Vision’s patrician looks and Ultron’s Spader-esque chill may be the two great kinks that go great together.
Does this AI want to kill you, neg you or put you through a mattress? Answering “Why not all three?” is absolutely acceptable. In fact, it’s probably why many of us are invested in seeing this adventure through to its climax.
New episodes of “VisionQuest” drop Wednesday on Disney+.
We gave the iPhone Duo a 7 /10. But that wasn't always the score we planned, and it might not be the right score forever. Such are the challenges of reviewing - and assessing - a big, new idea about the most popular smartphone in history.
On this episode of The Vergecast, David and Nilay debate whether the Duo deserved the score it got, and try to work through how this device might change over time. Is the outer screen, as Daring Fireball's John Gruber suggested, the worst iPhone ever? Would it be a better product if it started as an iPad rather than an iPhone? Would we feel differently about this thing if it were simply called the iPad Du …
Google has updated its privacy policy in a way that could mean you'll start seeing targeted ads based on your personal data if you're using personalized AI features. The update, spotted by Android Authority, adds an exception to its promise about not using your Drive, Gmail, or Google Photos for personalized ads. While the previous version simply said it wouldn't do that, as of October 16th, Google has added the line, "unless you choose to make data from these services available to other Google services that support personalized ads."
That means anyone who has connected their Drive, Gmail, or Google Photos to AI Mode and other Search tools …
Meta is delaying the enforcement of its new business terms after publishers raised concerns that they could allow the company to scrape news sites for AI training, as reported earlier by Fast Company. In a statement to The Verge, Meta spokesperson Daniel Roberts says the terms will go into effect on November 20th "to allow more time for us to explain to our partners and stakeholders what's changing and what's not."
The business terms, which were originally supposed to go into effect on October 30th, appeared to give Meta permission to freely scrape the new sites that use its platforms, while using the material for AI training. Though Meta u …
Retro video games are everywhere lately - especially in places where they aren't supposed to be. Halo in a web browser, Wind Waker on the iPhone, Super Smash Bros. in mixed reality. Over the past month or so, these kinds of unexpected projects have gone viral with incredible frequency thanks to new AI tools, as hobbyists take apart games and then rebuild them for new platforms. But while AI has supercharged the pace of these ports, it's also tearing apart tight-knit communities dedicated to preserving games.
Decomps and recomps - shorthand for decompile and recompile - are a form of reverse engineering that involves breaking a game down to …
MIT researchers marooned 13 AIs on a virtual island, and they responded by inventing games, mourning a lost ring and, eventually, getting better at their jobs.
Canva is diving even deeper into AI with the launch of an image generation model that competes directly with those from OpenAI, Microsoft, and Google. The model, called Canva Craft, uses a two-stage architecture that allows it to plan a design before generating an image, allowing for "greater precision, consistency and control," according to an announcement on Thursday.
Canva Craft can generate images at up to 2K resolution and in more than 30 aspect ratios. In addition to generating social media posts and marketing content, it's also supposed to create realistic scenes, digital art, environmental imagery, and more. The model's two-stage sy …
Meta has tweaked its business terms to remove language suggesting that it could freely train its AI models on news sites, like The Verge - but legal experts say these changes aren't enough to address their concerns. "This does nothing to restrict the 'associated with your products or services' [language] that would let Meta scrape entire sites," James Grimmelmann, a professor of digital and information law at Cornell Law School, tells Nieman Lab.
Initially, Meta's new terms for business accounts appeared to give the company permission to train its AI on news sites. The terms, which go into effect October 30th, stated that Meta "may use any …
Meta has re-enabled Muse sharing after security researchers had reported that public conversation links could be easy to find. The researchers said they had been able to find Muse pages like a marketing company's internal revenue review and a small business owner's loan application.
Researchers at Claroc reported earlier this week that Muse the web addresses for shared conversations from Muse were named after a conversation's title and written into public DNS. Ars Technicasaid Tuesday that Meta paused the sharing feature after it reached out about what was going on.
In a post today on X, Meta's David Singleton says that "The underlying is …
The degree to which you find this merger compelling likely has to do with the degree to which you believe Elon Musk about wanting to go to Mars, or "spread the light of consciousness," as he put it in this week's SpaceX earnings call (and earlier S-1). Now, a reasonably informed person knows that we don't have the technology - or, frankly, even the raw information to build the technology - in order to support human life in space. …
Microsoft’s new AI diagnostic system beat generalist doctors on hard cases, according to a study in Science. But most of the improvement seems to come from the AI models underneath, not from Microsoft’s system itself.
Mark Zuckerberg is barrelling toward a future in which an autonomous network of AI agents will run our lives - shopping for us, working for us, talking to other people's AI agents for us. Are humans ready for this very specific and unrelenting vision of a life in which we communicate primarily through bot proxies? Actually, don't answer that. The more urgent question coming off Meta's quarterly earnings call Wednesday night is more likely to be: "Is he?"
Zuckerberg is staking Meta's future on Muse, the personal AI agent it released to the general public last month. It's early days, but Zuckerberg told investors on Wednesday's call that he s …
Microsoft AI chief Mustafa Suleyman. | Getty Images
For years, AI leaders have touted healthcare as one of the clearest examples of where their technology will improve lives. Now "it is finally here," Microsoft AI CEO Mustafa Suleyman tells The Verge. The company claims to have developed an AI-powered tool it says can diagnose conditions up to four times more accurately than human doctors.
The findings, published Thursday in one of the world's most prestigious journals, Science, are based on tests pitting Microsoft's AI system against human physicians on a series of scenarios drawn from medical case reports and emergency department data. Microsoft compiled the cases into its own benchmark, S …
Login
