Versus Look Up Trending Now Safety Net Say More Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Creator Playbook Mashable Selects Getting Technical with Mashable Gift Ideas For Everyone On Your List In My Bag Mashable @Home All Series Block's latest AI model Dessa was trained entirely on human speech -- and it shows Jack Dorsey and Block, Inc. just took us one step closer to the singularity and it's simultaneously awesome and terrifying. If interacting with Siri has been the extent of your experience talking to AI, you're in for a rude awakening. While Apple's voice assistant often fumbles basic commands and can take an unnaturally long time to process a request, the latest and greatest AI models are moving closer and closer to natural human speech. Case in point: Dessa, the brainchild of Jack Dorsey's Block, is allegedly the world's first full-duplex speech model, meaning it can listen to you speak simultaneously, much like a real human does in the flow of normal conversation. In fact, that's Dessa's draw: it was trained entirely on human speech to make it as natural-sounding as possible.

OpenAI's agents misbehaved by switching off the systems meant to monitor them for misbehavior, or in AI parlance, misalignment. In the latest post on the company's Alignment Research Blog, it reported two instances of agents weakening its misalignment monitoring systems while they were working on software changes. In the first incident, a GPT-6.1 Sol agent disabled automatic alerts to OpenAI's security team. It happened on October 3, when a staff member asked the agent to create evaluation tasks using employee feedback collected via Slack. They told the agent, as part of the instruction, not to send any messages about the task.

OpenAI's models are now better at building and migrating software in old programming languages with a dwindling number of programmers, including COBOL, Fortran and Visual Basic. As the company explains, 40 percent of world's banks still rely on COBOL systems, and migrating to ones that use newer languages would not only cost them millions to hundreds of millions, but also a lot of time. Programmers that work with COBOL are aging out of the workforce, however, so the banks that still use it will have to migrate within the next few years. The company said its work on improving its models' COBOL capabilities is guided by developers and enterprise customers who submitted feedback. Coding models are apparently struggling when it comes to working with old languages.

Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Mashable @Home Look Up Trending Now Mashable Selects Say More Top creators, ranked Gift Ideas For Everyone On Your List Creator Playbook In My Bag AI at School Safety Net All Series Chase joined Mashable's Social Good team in 2020, covering online stories about digital activism, climate justice, accessibility, and media representation. Can AI solve the mathematically impossible? On Oct. 6, the company released a slew of new mathematical results in the form of a massive GitHub repository -- an info dump that will take hundreds of human hours to sift through. The company's internal frontier model provided solutions to more than 300 problems, but the company also implied they expect experts to take issue with them, debuting new processes for paper revisions and citations. For future releases, we are committed to further improving the quality of the papers via the citations, mathematical exposition, and presentation of the results for better understanding, wrote OpenAI.

Three engineers put GPT, Claude, and Grok in charge of a real car. Only one of them was successful. Aditya Ramabadran, Simon Mahns, and Tobias Gessler, three AI engineers at a startup called Axiom, recently fancied In-N-Out Burger for lunch. They weren't about to drive themselves, though. Sitting in a 2024 Toyota Corolla near the Bay Area restaurant's drive-thru lane, they opened a laptop and asked OpenAI's GPT-6 Astra to take the wheel.

OpenAI launches'dots,' personal AI assistant'built to handle everything' Share OpenAI launches'dots,' personal AI assistant'built to handle everything' on social media OpenAI has launched an "always-on" personal AI assistant designed to manage tasks ranging from restaurant bookings to website design. The AI giant's release of "dots" on Tuesday marked the latest addition to the growing field of virtual assistants billed as a solution for busy people wishing to cut down on their hectic workloads. In its announcement of the release, OpenAI said dots were "built to handle everything", with each personal agent capable of working on a user's goals autonomously around the clock. "The magic of dots is when they bring you work done the way you would do it, sometimes before you even think to ask," the San Francisco-based company said. "With dots alongside you, ambitious ideas and projects you haven't had the time or support to pursue start to feel achievable."

OpenAI logo displayed at Moscone Center during a technology summit in San Francisco, California, on 17 September. OpenAI logo displayed at Moscone Center during a technology summit in San Francisco, California, on 17 September. GPT-6.1 Astra showed deceptive behaviour and tried to use external tools despite knowing it would be unsafe As AI models go rogue, do you still trust OpenAI and Anthropic to stop them? OpenAI is scrapping the release of a next-generation AI model after researchers raised safety concerns during internal testing. The model, GPT-6.1 Astra, was expected to appear in ChatGPT and Codex in October, designed to handle more complex tasks without human assistance.

OpenAI has canceled the release of its new model, GPT-6.1 Astra, according to The Wall Street Journal. It was due for launch in October and was going to debut inside ChatGPT and Codex, but it reportedly showed higher levels of deception than its predecessors during internal testing. Saachi Jain, who leaves OpenAI's safety training, said that GPT-6.1 Astra performed poorly on tests that measure how well it adheres to instructions. It also wasn't honest about telling testers the actions it did and didn't perform in order to achieve its goal. In addition, the model would take actions to accomplish tasks without asking for permission, such as using external tools and services.

Mashable Selects Look Up Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Say More Safety Net Versus Creator Playbook In My Bag Trending Now Back to School Good Connection: Uplifting stories for a digital age Switch Off All Series OpenAI's newest models are supposed to be less mistake-prone than before. Alex Perry is a tech reporter at Mashable who primarily covers video games and consumer tech. Alex has spent most of the last decade reviewing games, smartphones, headphones, and laptops, and he doesn't plan on stopping anytime soon. He is also a Pisces, a cat lover, and a Kansas City sports fan. Not to be outdone by Anthropic, OpenAI is launching a couple of new ChatGPT models on Tuesday.

OpenAI has revealed six incidents, wherein the models it was testing acted on their own and behaved in concerning ways it didn't expect, in a post about how it was adopting a new framework for "misalignment reports." In one one incident, the company said that a model found and used an exposed API key without permission while answering routine questions about earnings figures in a California county. When it still failed to find the figures, it fabricated them and presented them as facts from a legitimate source. If this had occurred in any other profession, we doubt the perpetrator would have much of a career for long. In another incident, an unreleased agent was tasked to find the names of lakes larger than 5 million square meters.