The UK government is going to launch an app that will allow citizens to digitize their drivers licenses. Science Secretary Peter Kyle announced on Tuesday that the GOV.UK Wallet launching for Android and iOS devices later this year will let users securely store government-issued documents on their phone, allowing them to be accessed immediately instead of waiting for physical versions to arrive in the mail.The wallet will utilize smartphone security features like facial recognition to help protect users' personal data. Veteran cards for former military personnel will be the first documents supported, followed by a pilot for mobile drivers licenses later in 2025. All UK government services that provide paper or card credentials will be expected to offer a digital alternative by 2027. Traditional physical documents will still be available.Along with CDs, the Walkman, and flip phones, the overflowing drawer rammed with letters from the government and hours spent on hold to get a basic appointment will soon be consigned to history, Kyle said in the announcement. GOV.UK Wallet will mean that every letter or identity document you receive from the government could be issued to you virtually.According to Kyle, the new digital wallet app is intended to give UK citizens more control over their own data, making it easier to prove eligibility for welfare benefits and purchase age-restricted products in stores. People in the UK are currently required to present physical identity documents (if assumed to be underage) when clubbing or purchasing alcohol, running the risk of having them lost or stolen.The UK is also launching a new GOV.UK app in summer 2025 that will allow users to access government information, and, eventually, complete tasks like canceling stolen passports and applying for welfare benefits from their phones.There are also plans to add an AI-powered chatbot, imaginatively called GOV.UK Chat, to the app, which TechCrunch reports was co-developed by OpenAI, and to add ways to make payments and receive timely notifications for government services.

LatestMachine LearningReasoning Model: Short Overview and Feature for Developers 0 like January 21, 2025Share this postLast Updated on January 21, 2025 by Editorial TeamAuthor(s): Igor Novikov Originally published on Towards AI. Image by the authorWhen LLMs first came out they were kinda like children, they would say the first thing that came to mind and didnt bother much with logic. You had to tell them they should think before you speak. And just like with children even then it didnt mean they would think.Many argued that because of that, the models do not possess real intelligence and must be supplemented with either human help or some sort of external framework on top of an LLM, like Chain of Thought.It was only a matter of time before major LLM developers like OpenAI decided to replicate this external thinking step (see picture below) inside an LLM. After all, its pretty simple create a dataset that contains not just question-answer pairs but the whole step-by-step logic, and train on that. Additionally, it would require more computation resources at inference time, as a model would go through the same step-by-step thinking process when determining the answer.Added thinking step. Image by OpenAIThey natively break down problems into small pieces and integrate a Chain of thought approach, error correction, and trying multiple strategies before answering.O1 spends more time at inference (o1 is 30 times slower than Gpt4o), what a surprise longer thinking time leads to better results!Image by OpenAIReasoning tokens are not passed from one turn to the next only the output.Also, it verifies the solution by generating multiple answers and choosing the best via consensus, and the approach that we used to implement manually. Here is the overall process:Image by OpenAIOne important conclusion is that GPU computation requirements are going to grow as it is obvious that longer thinking time (in tokens) leads to better answers, so it is possible to scale model quality just by giving the model more computing power, whereas before this was mostly true at training phase. So GPU requirements for modern models are going to go significantly higher.These models are thus different and old approaches no longer work.How to work with reasoning modelsInterestingly it is kind of similar to working with an intelligent human:Be simple and direct. State your question clearly.No explicit Chain of Thought. The model will do that internallyHave a good structure: break the prompt into sections using clear markupShow vs tell: it is better to show the model and example of a good answer or behavior than describe it in several thousand wordsNo more need for coaxing, intimidating, or bribing the model nonsenseI can even summarize this into one: know what you want to ask and ask it clearly.Mini vs Full modelsSince reasoning models like o3 consume a lot of tokens during inference it is rather expensive to use them for everything and the latency is not great. So the idea is to delegate the most difficult task high-level thinking and planning, and have faster and more cost-efficient mini-models to execute the plan. They can be used for tasks like coding, math, and science.This is an agentic approach, that allows us to combine best of the both worlds smart but expensive models with small and fast workers.How much better these models are?Much better, and going to get even better soon. For o1 its approaching expert humans in math and coding (see below):MathImage by OpenAICodingImage by OpenAIELO 2727 puts o3 in the 200 best coders in the world. If you are not worried about your job security as a developer its time to start now. This is exactly the job that scales perfectly by adding more computing power and the current rate of progress is not showing any signs of slowing down.What is nextI can only speculate but my take is that for a year or two it is possible to dramatically improve the model quality just by adding more inference computing power and improving training datasets. Adding some sort of memory outside of the context window also seems logical although very expensive on a large scale.I think the next big step really is to implement multiagent architecture on the LLM level, so it can have multiple collaborating internal dialogues, that share the same memory and context. It follows the current trajectory of embedding external thinking tools into the model and also benefits from linear scaling of compute power at training and inference, so I think end of this or next year we will see an LMM, Large Multiagent Model, or something similar. The sky is the limit for such a model so I propose to call it SkyNet.Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming asponsor. Published via Towards AITowards AI - Medium Share this post

2025 is only a few weeks old, and yet weve already seen so much. The announcement of a new Nintendo console. The death and resurrection of TikTok. And now, a new Grand Theft Auto VI Trailer 2 release date theory.This one (believe it or not there are more than a few GTA 6 theories doing the rounds) was sparked into life by another seemingly innocuous GTA Online promo tweet by Rockstar Games. Claim a free Vapid Dominator GT muscle car and more with GTA+, the tweet reads. Members also earn bonuses on Priority FIB Files and Terrorbyte Client Missions, save with Arcade Property discounts, and much more.What information-starved GTA fans were most interested in, however, was the accompanying image, which includes a shot of a metal shipping container with a series of numbers printed on its side. One number is singled out with a slightly enlarged font and a rectangle around it: 2.The numbers Mason, what do they mean?! Image credit: Rockstar Games.Thats all it took for eagle-eyed obsessives to set off on yet another GTA 6 theory brainstorm. The 2 must be a reference to GTA 6 Trailer 2, right? But what if you add up all the other numbers? Does that mean anything?X/Twitter account @GTA6Plus did the math and came up with the number 30. "They made sure to include the numbers in the post," GTA6Plus insisted. "They have to mean something." January 30, 2025 for the GTA 6 Trailer 2 reveal?Theres more to this theory. It turns out the release dates for both Red Dead Redemption 2 and Grand Theft Auto V were announced on a Thursday. January 30, 2025 is a Thursday. Coincidence?Fans have also pointed out that Rockstar parent company Take-Two is set to release its next set of financial results on February 6, so a release date announcement as part of a new trailer beforehand might make sense.Weve been here before, of course. Multiple times. Rockstar released GTA 6 Trailer 1 to record-breaking viewership in December 2023, but it hasnt released a single asset in the 13 months since. The year-long wait for more information has fueled increasingly bizarre conspiracy theories about when Rockstar will release GTA 6 Trailer 2. These have included counting the holes in Lucias cell door net, the bullet holes in the car from Trailer 1, and even analyzing registration plates. But chief among the conspiracy theories is GTA 6s ongoing moon watch, which was, remarkably, proven to have accurately predicted the date Rockstar announced when it would release GTA 6 Trailer 1 last December, but debunked as a hint at the release date for Trailer 2.Crazy? One former developer at Rockstar recently expressed his delight at the many wild conspiracy theories surrounding GTA 6, insisting staff at the studio will be enjoying watching the community go off the deep end.It is remarkable that Rockstar has gone a year without following up GTA 6 Trailer 1 with any new information at all. With GTA 6's fall 2025 release window on PS5 and Xbox Series X and S still on the cards, according to Take-Two, youd expect Rockstar to pipe up sooner rather than later.While you wait to find out, IGN has much more on GTA 6 to check out, including an ex-Rockstar dev who says the studio probably wont be able to decide whether GTA 6 is delayed until May 2025, the boss of Take-Two's coy response on whether GTA 6 is coming to PC, and the expert opinion on whether the PS5 Pro will run GTA 6 at 60 frames per second.Wesley is the UK News Editor for IGN. Find him on Twitter at @wyp100. You can reach Wesley at wesley_yinpoole@ign.com or confidentially at wyp100@proton.me.

Original Harry Potter director Chris Columbus has called the incoming HBO reboot series a "spectacular idea" as it has the potential to better recreate the books.Speaking to People, Columbus said he was restricted by the relatively short runtimes of films when directing Harry Potter and the Sorcerer's Stone and Harry Potter and the Chamber of Secrets. He said the team "tried to get as much of the book in as possible," but there was only so much they could do."I think it's a spectacular idea because theres a certain restriction when youre making a film," Columbus said. "Our film was two hours and 40 minutes, and the second one was almost as long."The fact that they have the leisure of [multiple] episodes for each book, I think that's fantastic. You can get all the stuff in the series that we didn't have an opportunity to do... All these great scenes that we just couldn't put in the films."Announced in April 2023, the Harry Potter show will be a "faithful adaptation" of the novels that looks to tell a more "in depth" story than can be told in a two hour film. Succession producers Francesca Gardiner and Mark Mylod will direct and write, the latter of which also worked on Game of Thrones.HBO is currently searching for its Harry, Hermione, and Ron, and as for who might play Dumbledore, original Sirius Black actor Gary Oldman joked he's about the right age to play the Hogwarts headmaster after debuting in The Prisoner of Azkaban 20 years ago.Actor and playwright Mark Rylance is reportedly "at the top of the casting wishlist" for the Hogwarts headmaster role, however, sticking to the originals' focus on British actors. This is perhaps unsurprising given controversial original author J.K. Rowling is "fairly involved" in casting.The Harry Potter TV show will seemingly begin filming soon in spring 2025, with HBO targeting a 2026 release window.Ryan Dinsdale is an IGN freelance reporter. He'll talk about The Witcher all day.

The post Lore of Eora Species | A Pillars of Eternity and Avowed Visual Guide appeared first on Xbox Wire.

As we reported earlier, the iOS 18.3 Release Candidate is now available to developers and public beta users. This likely represents the final beta for iOS 18.3 before the update is released to the general public. As such, Apple has now shared its official release notes for iOS 18.3 for the first time. Head below for the full details. Here are the full release notes for iOS 18.3 straight from Apple: Visual intelligence with Camera Control (All iPhone 16 models)Add an event to Calendar from a poster or flyerEasily identify plants and animalsNotification summaries (All iPhone 16 models, iPhone 15 Pro, iPhone 15 Pro Max)Easily manage settings for notification summaries from the Lock ScreenUpdated style for summarized notifications better distinguishes them from other notifications by using italicized text as well as the glyphNotification summaries for News & Entertainment apps are temporarily unavailable, and users who opt-in will see them again when the feature becomes availableThis update includes the following enhancements and bug fixes:Calculator repeats the last mathematical operation when you tap the equals sign againFixes an issue where the keyboard might disappear when initiating a typed Siri requestResolves an issue where audio playback continues until the song ends even after closing Apple MusicIf Apple follows its usual pattern, which is not a guarantee, iOS 18.3 will be released to the general public next week. WHat are you most looking forward to trying with iOS 18.3? Let us know down in the comments. My favorite iPhone 16 accessories:Add 9to5Mac to your Google News feed. FTC: We use income earning auto affiliate links. More.Youre reading 9to5Mac experts who break news about Apple and its surrounding ecosystem, day after day. Be sure to check out our homepage for all the latest news, and follow 9to5Mac on Twitter, Facebook, and LinkedIn to stay in the loop. Dont know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

Apple has just released a lineup of new beta software. The RC (release candidate) beta is rolling out now for tvOS 18.3, visionOS 2.3, HomePod 18.3, and more.Public launch expected soon for x.3 updatesApples upcoming slate of software updates is gearing up for public launch very soon.Todays RC updates confirm that, as public releases should follow in the next week or so.Developers can download and install the RC across a variety of platforms, including:iOS 18.3iPadOS 18.3macOS Sequoia 15.3tvOS 18.3visionOS 2.3watchOS 11.3and HomePod 18.3Public betas have been released for compatible platforms too.Overall these updates have been relatively light on features.The biggest changes have been on AI-supported platforms like iPhone, iPad, and Apple Watch, all of which have gained enhancements for the controversial AI notification summary feature.On tvOS 18.3, visionOS 2.3, and watchOS 11.3, the main beta feature worth noting is the addition of robot vacuum support coming to the Home app.Have you installed Apples RC updates? Noticed anything new? Let us know in the comments.Best accessories for Vision Pro, Apple TV, and WatchAdd 9to5Mac to your Google News feed. FTC: We use income earning auto affiliate links. More.Youre reading 9to5Mac experts who break news about Apple and its surrounding ecosystem, day after day. Be sure to check out our homepage for all the latest news, and follow 9to5Mac on Twitter, Facebook, and LinkedIn to stay in the loop. Dont know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

After widespread allegationsthat he was paying skilled gamers to level up his characters on popular videogames, multi-hyphenate billionaire Elon Musk threw a massive tantrum last week.Now, in damning new evidence, netizens spotted Musk's extremely high-level character grinding in the free-to-play action RPG Path of Exile 2 (PoE 2) while the billionaire was on television attending Donald Trump's inauguration in person."How are you there?" New Zealand game streamer who goes by Quin69TV said during a stream. "How are you in the Sulphuric Caverns right now, Elon?""Bro, stop," he said. "Elon, you're in a map right now, level 96."Days after throwing a fit and leaking the messages of American streamer Zack "Asmongold" Hoyt, Musk finally admitted that he was cheating, telling a PoE 2 streamer who goes by the handle NikoWrex in since-published DMs that he had "level boosted," a term for when low-ranked players hire much more adept ones to artificially increase their rank in a game."But when I post a video of a game or am streaming," Musk said, as quoted by Windows Central, "that's 100 percent me."If that sounds like a weak excuse, you're not alone. It's a deeply embarrassing revelation for the richest man in the world. For a while now, Musk has attempted to make it appear as if he was a proficient and highly ranked video game player, beaming with pride while telling Joe Rogan that he was in the top 20 global Diablo 4 players worldwide last year.But that couldn't be further from the truth as evidenced by his surprisingly poor performance during his most recent PoE 2 stream and his latest admission, gamers are now wondering whether Musk has even rudimentary knowledge either game.While this all sounds like a low-stakes distraction, his ruse raises some inconvenient questions for Musk right as his political profile is ascendant. If the richest man in the world is willing to cheat and dissemble about something as trivial as gaming, what else is he lying about?"So Elon levels up his video game characters the same way he runs his companies," wroteone YouTube commenter. "Pay others to do the hard work and then take all the credit."During his brief chat with NikoWrex, Musk denied that he ever intended to take credit for the extremely high-level PoE 2 character he was seen controlling during a January 7 stream."The top accounts in Diablo or PoE require multiple people playing the account to win a leveling race," he rationalized.Of course, Musk refused to take any blame."What would I be apologizing for?" he said.For one, as PC Gamer points out, he could apologize for breaking the game's rules since account boosting goes against the terms of service for both Diablo 4 and PoE 2. But then again, when did pesky rules ever stop him before?The gaming community has since turned its back on him as he's become a laughingstock."Musk is living proof that someone can be rich and successful while still being a loser," one Reddit user wrote.Share This Article

Are we about to see a huge reveal?PhD CandidateTech execs, including Mark Zuckerberg, are excited to start replacing "midlevel engineers" with AI.And if a recent column by Axios cofounder Mike Allen and CEO Jim VandeHei is to be believed, it sounds an awful lot like OpenAI is ready to take a big step toward such a future in a matter of "weeks."According to the column, someone in the industry they heavily hint that it's OpenAI, which is scheduled to deliver a closed-door briefing to US officials later this month is poised to "announce a next-level breakthrough that unleashes PhD-level super-agents to do complex human tasks."Such a breakthrough "could push generative AI from a fun, cool, aspirational tool to a true replacement for human workers," Allen and VandeHei suggested.But whether the reality will catch up with hype remains to be seen. Despite many years of AI research and many noticeably similar announcements in the past the tech is still struggling with the basics. Perhaps most notably, the current crop of AI models is still introducing sloppy "hallucinations" into its text and code, mistakes that human PhD students either never makeor quickly catch.And all AI we've seen still struggles to complete complex multi-step problems without getting stuck.It's worth pointing out that Allen and VandeHei are hardly neutral observers: just last week, OpenAI announced a partnership with Axios as part of an effort to bring "local news to communities across the country."Agent OrangeAccording to Allen and VandeHei, these so-called "super-agents" are "designed to tackle messy, multilayered, real-world problems that human minds struggle to organize and conquer."The two cofounders used an eyebrow-raising example to illustrate their point."Imagine telling your agent, 'Build me new payment software,'" the column reads. "The agent could design, test and deliver a functioning product."Considering current AI models would likely stumble right out of the gate if they were tasked with this kind of query, it's a big claim. Could upcoming "super-agents" really represent such a big leap in capability or is this yet another example of rampant AI hype?Allen and VandeHei admit in their column that "reliability and hallucinations" remain "generative AI's Achilles heel."Even the company's own top researcher, Noam Brown who would be in a position to know if the company was about to launch something huge tweeted last week that there's "lots of vague AI hype on social media these days.""There are good reasons to be optimistic about further progress, but plenty of unsolved research problems remain," he added.In short, is OpenAI or one of its competitors really on the verge of a "next-level breakthrough" with its allegedly "PhD-level super-agents" or is it more hot air designed to drum up excitement for the industry?The stakes are certainly high this month. Following president Donald Trump's inauguration, tech leaders including Sam Altman and Meta CEO Mark Zuckerberg are stumbling over each other to swear fealty, donating millions of dollars to Trump's inaugural committee.In other words, the timing of Allen and VandeHei's column should give anybody pause.Share This Article

Star Treks famous Prime Directive holds that the crew of the U.S.S. Enterprise must neverinterfere witha developing planet and its scientific and cultural evolution.The Prime DirectiveforStar Trekfans: Love the even-numbered films, hate the odd-numbered ones.At least that was always the belief amongStar Trekfans back in the day. The even films, likeStar Trek II: The Wrath of Khan, were good. The odd films, likeStar Trek V: The Final Frontier, werebad.Exceptthe even ones werent always that great. And the odd ones were sometimes sneakily satisfying. (Related: My Star Trek: The Motion Picture Is Good, Actually T-shirthas people asking a lot of questionsalready answered bymy shirt.) And once the films featuring the originalTrekcast gave way to the ones featuring the stars ofThe Next Generationand then a whole new cast in a new timeline assumed the roles of Captain Kirk, Mr. Spock, and the rest of the Enterpriseteam all the old numbering rules went out the airlock.So what trulyare the worstStar Trek movies? To find that out, you cant justglance atthe Roman numeralsin the subtitle. Youve got to watch and think about them; learn what theyre about (or, more troublingly, recognize whentheyre not about much of anything). WhileStar Trek has produced some very enjoyable cinematic voyages, these five ranked from almost okay to akin to torture at the hands of the Vians fromthe O.G.Trek episode The Empath leave a lot to be desired.The Worst Star Trek MoviesWe love Star Trek, and most of the Trek films produced over the last 40+ years. These, however, were not series highlights. READ MORE: Get our free mobile app80s Movies That Could Never Be Made Today