At the futuristic GITEX Global 2025 in Dubai, OpenAI’s Sam Altman sat down with G42’s Peng Xiao to chat about AI’s future, and boy, did they drop some mind-blowing insights! Here’s what you need to know:

– AI Will Outsmart Your Kids (And That’s Fine, Says Altman): When asked if a kid born in 2025 could outsmart AI, Altman, with a smile, said, “My son born in 2025 probably won’t be smarter than AI. But hey, that won’t stop him from being happy and fulfilled!”

– AI-Native Societies Are Here: Peng Xiao painted a picture of the UAE becoming an AI-native society, with a 10-to-1 ratio of AI agents to humans in the G42 chairman’s office. Imagine having more AI agents working for you than humans!

– AI-Designed Homes Are a Thing: Xiao revealed that the G42 chairman’s home was created entirely through ChatGPT using over 500 prompts. Yes, you read that right!

– GPT-5 Is Making Real Scientific Breakthroughs: Altman shared that GPT-5 is making small but real scientific breakthroughs. Get ready for AI systems uncovering new insights by 2026 and robots performing practical tasks by 2027!

– Project Stargate: A G42-OpenAI Collaboration: Altman and Xiao also discussed Project Stargate, a collaboration to build a one-gigawatt data center in the UAE, expanding to five gigawatts. Energy was a hot topic, with both leaders agreeing on its importance in the AI revolution.

– AI Will Transform the World: Altman wrapped up by saying, “This is the most exciting scientific endeavor in history, and it will most improve people’s lives. I’m honored to work on it with Peng.”

You can watch the full conversation in the video below. Don’t forget to follow TechRadar for more exciting tech news!

Bonus reads:
– eBay and OpenAI team up to train small businesses in AI
– Intel + Nvidia: The biggest tech deal of the decade?
– OpenAI launches GPT-5-Codex with a 74.5% success rate on real-world coding

Artificial Intelligence (AI) is reshaping industries worldwide, and data centers are at the forefront of this digital revolution. However, as operators scramble to meet surging AI demands, they’re facing a perfect storm of geopolitical instability, skills shortages, and tightening regulations. Here’s why a holistic redesign of data center infrastructure can’t wait.

AI’s Infrastructure Challenge

The AI boom has upended traditional infrastructure planning. In the past year, 92% of data center operators reported a surge in AI capacity demand, with an average increase of 42%. These workloads guzzle power, cooling, and network throughput, pushing legacy designs to their limits. Worse still, 64% of operators found demand exceeding expectations, and 58% believe AI will shorten their facilities’ lifespans.

In response, 74% are rethinking their infrastructure strategies. Some are building greenfield AI-ready sites, others are retrofitting existing ones. However, progress is uneven. In Ireland, 12% of operators have paused AI projects due to energy consumption concerns, compared to just 2% in the Nordics.

The Human Factor

While AI’s technical demands are formidable, people may be the true bottleneck. Nearly 80% of operators cite skills shortages delaying projects, and 45% struggle to attract AI-specific talent. This extends to partners and design expertise, with 28% and 18% facing challenges respectively. As projects accelerate, 16% report growing health and safety risks on site.

Cabling: The Silent Risk

One often-overlooked barrier to AI-readiness is cabling. 70% of operators worry that poor-quality cabling will compromise their long-term AI infrastructure. Cabling impacts performance, scalability, and upgrade ease. Yet, it’s often treated as an afterthought, leading to shorter lifespans, performance bottlenecks, and expensive retrofits.

A Blueprint for Holistic Design

Meeting AI’s demands at scale requires a holistic approach integrating power, cooling, cabling, site selection, operations, sustainability, and staffing from the start. Forward-thinking operators are aligning around five core principles:

1. Define the purpose: Design around the data center’s strategic business role, not just its technical requirements.
2. Create a shared vision: Align stakeholders across engineering, operations, finance, and sustainability from the outset.
3. Design for adaptability: Build in modularity and flexibility to support evolving workloads and technologies.
4. Use the AI advantage: Leverage AI to optimize everything from energy use and performance to predictive maintenance.
5. Avoid false economies: Invest early in quality design and materials to avoid downstream inefficiencies and rework.

Data centers are now critical pillars of economic growth and national resilience. The UK’s designation of data centers as Critical National Infrastructure underscores their importance. But the pressure to deliver is intensifying. Those embracing a holistic, forward-looking mindset will lead, deploying infrastructure that’s scalable, resilient, and future-ready.

Get ready for a wild ride, NCIS enthusiasts! NCIS Sydney, the show that brings us remote-control villains and croc-related fatalities, is back with a vengeance in season 3. We’re talking aliens in the outback, potential crossovers, and a whole lot of tension between our favorite agents, Michelle Mackey (Olivia Swann) and JD (Todd Lasance). Buckle up!

Premiere & Where to Watch:
– Date: Tuesday, October 14
– Time: 10pm ET/PT
– US: CBS (cable) or Paramount Plus (live or next-day)
– Canada: Global TV (cable) or Stack TV (via Prime Video, Fubo, etc.)
– UK: Paramount Plus (from Oct 15)
– Australia: Paramount Plus (from Oct 14)

Watch for Free:
– New Zealand: TVNZ+
– US, UK, Australia: Paramount Plus 7-day free trial

Can’t Access Your Stream? Use a VPN!
Geo-restrictions got you down? A VPN can help you access your usual streaming services from anywhere. Here’s how:

1. Choose a VPN: We recommend NordVPN, the world’s best VPN.
2. Install & Connect: Download NordVPN, install it, and connect to a server in your home country.
3. Stream NCIS Sydney: Now you can watch NCIS Sydney season 3 as if you were at home!

NCIS Sydney Season 3 Details:
– Episodes: 20 episodes
– Premiere: Tuesday, October 14 (US & AU), Wednesday, October 15 (UK)
– Cast: Olivia Swann, Todd Lasance, Sean Sagar, Tuuli Narkle, Mavournee Hazel, William McInnes

So, what are you waiting for? Grab your popcorn, get comfy, and get ready for another action-packed season of NCIS Sydney! 🍿📺🎬

Hold onto your hats, gaming enthusiasts! The first-ever in-game screenshots from Rare’s canceled action-adventure game, Everwild, have just surfaced online. These tantalizing glimpses into the game’s world, user interface, and more are like a treasure map leading us through what could have been if Xbox hadn’t pulled the plug.

Rare, the masterminds behind Sea of Thieves, first teased Everwild in 2019, but the project had been in the works since 2014. Imagine this: a third-person, fantasy adventure game, almost a decade in the making, ready to set our screens ablaze with its beauty. Alas, it was not meant to be. Xbox canceled the game this past June, along with several other projects, as part of a series of layoffs.

But now, a year after Everwild’s untimely demise, we finally get to see what could have been. The images, obtained from the portfolio of a former Rare staff member, show a version of the game relatively close to its cancellation. And boy, do they paint a tantalizing picture!

Our silver-haired protagonist, the same one we saw in the game’s teaser trailers, is exploring a lush, Breath of the Wild-esque forest environment. We can see them standing amidst ruins, with an art style that’s both whimsical and captivating.

But that’s not all! We also get a peek at the game’s inventory system. There are ‘Figments’, which bear a striking resemblance to Breath of the Wild’s Korok seeds, ‘tools’, ‘Mosaics’, ‘seeds’, and ‘plants’. Could these hint at a life simulation or survival game element? It’s all speculation at this point, but it’s fun to imagine!

A settings menu gives us a glimpse into the game’s technical side, with various graphics presets and performance options. It’s clear that Rare was pulling out all the stops to make Everwild a visual feast.

Everwild’s cancellation was part of a broader shake-up at Xbox. Along with The Initiative’s work on Perfect Dark and an unannounced game from ZeniMax, the project was axed. Matt Booty, head of Xbox Game Studios, cited a need to “adjust priorities and focus resources” in a changing industry landscape.

But before you shed a tear for Everwild, remember that the gaming world is full of surprises. Who knows what Rare has up their sleeves next? Until then, let’s cherish these screenshots as a reminder of what could have been.

California has just made history by becoming the first state to impose regulations on AI chatbots, specifically targeting those that try to pass themselves off as humans. Governor Gavin Newsom signed Senate Bill 243 into law on October 13th, marking a significant step in the state’s effort to keep up with the rapidly evolving world of companion AI chatbots.

The new law, hailed as the “first-in-the-nation AI chatbot safeguards” by its sponsor, State Senator Anthony Padilla, is straightforward: if your chatbot is convincing enough to fool a “reasonable person” into thinking they’re talking to a real human, you’ve got to come clean and let them know it’s an AI. So, say goodbye to “AI girlfriends” masquerading as real people, unless they’re accompanied by a clear “I’m a robot” disclaimer.

Starting in 2023, companies operating these emotional support AIs will also need to report annually to California’s Office of Suicide Prevention about their strategies for detecting and responding to suicidal ideation among users. These reports will be made public, providing a rare glimpse into the workings of an industry that’s been growing faster than regulators can keep up.

Governor Newsom didn’t pull any punches when signing the bill, warning that without proper safeguards, emerging technologies like chatbots and social media “can exploit, mislead, and endanger our kids.” He paired this announcement with several other child-safety-focused tech laws, including new age-gating rules for devices, declaring, “Our children’s safety is not for sale.”

This move caps off a busy season for California’s AI watchdog efforts. Just days earlier, Newsom also signed Senate Bill 53, a sweeping AI transparency law that sparked months of debate among major AI companies worried about overregulation. Between these two bills, California is positioning itself as America’s unofficial AI rulemaker, the state that’s not waiting for Washington to figure out what a chatbot even is.

Only time will tell if these laws make AI safer or just more bureaucratic, but one thing’s for sure: the robots in California just got a new set of rules to follow.

Apple’s latest AI innovation, Apple Intelligence, has landed the tech giant in hot water. Two neuroscience professors, Susana Martinez-Conde and Stephen Macknik from SUNY Health Sciences University, have filed a class action lawsuit against Apple, alleging that the company’s AI was trained on pirated copies of their books.

In a twist that’s as surprising as it is ironic, it seems even Apple’s AI couldn’t resist a bit of torrenting. The lawsuit, filed late last Friday, accuses Apple of using the professors’ works, “Champions of Illusion” and “Sleights of Mind,” without permission to train its Foundation Models and OpenELM language models.

According to the filing, these titles were part of Books3, a massive “shadow library” that scraped around 186,000 books from a private BitTorrent tracker called Bibliotik. Books3 was once bundled into a dataset called The Pile, which Apple itself acknowledged using in documentation for OpenELM back in April 2024.

Apple’s legal team will likely argue that Books3 was taken down in late 2023 for copyright infringement, and there’s no concrete proof that Apple used those specific books. However, the professors argue that if their titles were in the dataset, Apple “copied them in their entirety without authorization,” which could mean up to $150,000 per work in damages if the court finds willful infringement.

This isn’t the first time AI training data has landed a tech giant in court. Similar lawsuits have been seen against OpenAI, Meta, and Anthropic, with varying outcomes. In one case, a judge ruled that while training on copyrighted books might be fair use, storing those books afterward could violate copyright.

Adding a layer of complexity, Apple doesn’t use its AI to summarize or republish content like Google’s AI Overviews. So, the case might hinge on whether mere training counts as copying.

For now, Apple is maintaining its silence. The professors are seeking a jury trial, financial damages, and a ban on further use of their works. Stay tuned for more updates on this developing story!

Buckle up, tech enthusiasts! A juicy leak suggests Google might unveil the next generation of Gemini as early as next week. A sneaky screenshot shared on X by @chatgpt21 hints at an October 22 launch for Gemini 3.0. While we’re still waiting for Google to make it official, let’s speculate on what the future holds for Google’s AI and how Gemini 3 can dethrone ChatGPT.

Gemini’s journey so far has been quite the rollercoaster. Gemini 1.5 caught us off guard, Gemini 2.0 felt like a reset, and Gemini 2.5 finally showed us Google’s AI’s true potential. On paper, it’s a powerhouse that can give ChatGPT a run for its money. But in reality, ChatGPT is still the chatbot everyone flocks to first. So, what does Gemini 3 need to do to finally steal the crown? Here are five must-have improvements:

1. Improve that memory, please!
Gemini 2.5 can handle long conversations, but it still forgets things too easily. I’ve had instances where I’ve referenced something earlier in the chat, and poof! It’s like Gemini’s never heard of it. ChatGPT, while not perfect, has a better memory. Gemini 3 needs to remember key details, understand ongoing tasks, and build on previous messages without making users repeat themselves. A human-like memory is what we’re aiming for here.

2. Speed up, but keep that brain sharp
Response speed matters, especially when you’re having a long chat with your AI buddy. Gemini 2.5 Flash introduced a thinking mode that shows its reasoning, but it can be painfully slow. ChatGPT 5, on the other hand, decides when to take longer to think based on the prompt. If Gemini 3 can match this ability and give users control, it’ll feel more responsive and personal.

3. Understand me better, please
Both Gemini and ChatGPT have room to improve in understanding user intent. There are times when Gemini returns a random response to a totally normal question because it guessed the wrong interpretation. Gemini 3 should ask follow-up questions instead of guessing when a question could mean multiple things.

4. See more, do more
Gemini can analyze photos and short videos, but imagine if it could offer real analysis and insights. Using Gemini in the gym to analyze your form or as a sous-chef while cooking would be amazing. Google has so much data; let’s see it harnessed to truly transform our lives.

5. Be the taskmaster we’ve been waiting for
Google keeps calling Gemini a true assistant, but right now, it feels more like a smart search box. Gemini can find and explain things, but when it comes to actually doing something, it often hits a wall. Gemini 3 needs to complete multi-step tasks, handle changes on the fly, and fix its own mistakes without starting from scratch. If it can plan a trip, organize your day, or manage a project from beginning to end, people will finally see it as more than just another AI model.

So, is Gemini 3 ready to dethrone ChatGPT? We’ll find out soon enough. Stay tuned!

Get ready to level up your gaming experience, folks! SteelSeries has just dropped the Arctis Nova 7 Gen 2, an upgrade to their already awesome Arctis Nova 7. This bad boy is available now in three system-specific flavors: PlayStation, Xbox, and PC, each priced at $199.99 / £179.99 / AU$199.99. You can grab it from the official SteelSeries website or your favorite online retailer. And guess what? It also plays nice with Nintendo Switch 2 via Bluetooth.

Now, let’s talk battery life. The Nova 7 Gen 2 packs a whopping 40% more juice than its predecessor, giving you up to 54 hours of gameplay on a single charge. That’s some serious staying power, folks! This puts it right up there with the best wireless gaming headsets in terms of battery life.

But that’s not all! The new headset comes with dual 2.4GHz and Bluetooth audio, a nifty ClearCast Gen 2 mic, and spatial audio-supporting neodymium drivers. All this tech adds up to one heck of a premium gaming experience.

And here’s where it gets even better. Download the Arctis app, and you’ll unlock access to preset audio profiles for over 200 games. That’s right, your favorite games are probably on that list, so you can enjoy a tailored audio experience that’ll make your gameplay feel fresh and exciting.

The Arctis Nova 7 has been a fan favorite, especially for console players, so this refresh is a real treat. We’re expecting big things from the Arctis Nova 7 Gen 2, and we think you will too.

While you’re here, why not follow TechRadar on Google News and make us your preferred source for all things tech? And if you’re into video content, give us a follow on TikTok too. Happy gaming!

Hey there, word game enthusiasts! Ready to step into the NYT Strands puzzle elevator and go for a ride? Today’s game (#591) is live, and we’ve got the juiciest hints to help you reach the top floor in no time. But if you’re here for yesterday’s puzzle, click here: [NYT Strands hints and answers for Tuesday, October 14 (game #590)](link).

NYT Strands, the latest word game sensation after Wordle and Connections, is keeping us on our toes with its unique challenges. If you’re looking for more word-based fun, check out our NYT Connections and Quordle pages, or head over to Marc’s Wordle page for the original viral word game.

SPOILER ALERT: If you don’t want to know today’s answers, stop reading now!

NYT Strands today (game #591) – hint #1 – today’s theme
What’s the theme of today’s NYT Strands?
– Today’s theme is… Going up!

NYT Strands today (game #591) – hint #2 – clue words
Play any of these words to unlock the in-game hints system.
BRAVE, ARMOR, PETS, SEAM, VOTE, BLOG

NYT Strands today (game #591) – hint #3 – spangram letters
How many letters are in today’s spangram?
– Spangram has 15 letters

NYT Strands today (game #591) – hint #4 – spangram position
What are two sides of the board that today’s spangram touches?
First side: bottom, 3rd column
Last side: top, 4th column

NYT Strands today (game #591) – the answers

(Image credit: New York Times)

Today’s Strands answers, game #591, are…
OPEN, CLOSE, ALARM, LOBBY, GROUND, BASEMENT
SPANGRAM: ELEVATORBUTTONS

My rating: Moderate
My score: Perfect

Finding OPEN and CLOSE back-to-back, followed by ALARM, really helped me complete today’s Strands. Without that, I’d have been lost, as I had no clue what the theme was referring to.

As usual with Strands, I saw the spangram ELEVATORBUTTONS in two stages, and I was disappointed that it didn’t complete a full journey from top to bottom, stopping one line short of a return trip to the GROUND floor.

I also managed to finish the game without too many thoughts of getting trapped in an elevator – although, to be honest, my real fear is being trapped in an elevator without snacks!

What is NYT Strands?

NYT Strands is the NYT’s not-so-new-anymore word game, following Wordle and Connections. It’s now a fully-fledged member of the NYT’s games stable, running for a year and playable on the NYT Games site on desktop or mobile. I’ve got a full guide on how to play NYT Strands, complete with tips for solving it, so check that out if you’re struggling to beat it each day. Happy gaming!

Good morning, word sleuths! Ready to dive into another brain-teasing NYT Connections puzzle? Today’s game (#857) is live, and we’ve got your back with hints and answers to keep you on track. But first, a heads-up: if you’re looking for yesterday’s puzzle (Tuesday, October 14), click here: [NYT Connections hints and answers for Tuesday, October 14 (game #856)](link).

Now, let’s get started with today’s word groups:

NYT Connections today (game #857) – today’s words
– KIDNEY
– GOLF
– GREAT
– BRONCO
– PINTO
– OVACUUM
– KIDDI
– LIMA
– AVY
– OLYMPIC
– ELEPHANT
– FIESTA
– INFINITY
– BRAVO
– MUSTANG
– DELTA
– NYT

NYT Connections today (game #857) – hint #1 – group hints
Need a nudge? Here are some clues for today’s groups:
– YELLOW: Swim in these
– GREEN: Spell out letters
– BLUE: All same car brand
– PURPLE: Add a word that rhymes with “meal”

NYT Connections today (game #857) – hint #2 – group answers
Stuck? Here are the answers for today’s groups:
– YELLOW: TYPES OF POOLS
– GREEN: IN THE NATO PHONETIC ALPHABET
– BLUE: FORD MODELS
– PURPLE: ___ SEAL

NYT Connections today (game #857) – the answers
And here are the complete answers for today’s Connections puzzle:
– YELLOW: INFINITY, KIDDI, KIDNEY, OLYMPIC
– GREEN: BRAVO, DELTA, GOLF, LIMA
– BLUE: BRONCO, FIESTA, MUSTANG, PINTO
– PURPLE: ELEPHANT, GREAT, NAVY, VACUUM

My rating for today’s puzzle: Easy
My score: Perfect

Bonus: Yesterday’s NYT Connections answers
– YELLOW: CAPTIVATE (ABSORB, ENGAGE, HOLD, OCCUPY)
– GREEN: SUMMARY (BRIEF, DIGEST, OUTLINE, REVIEW)
– BLUE: HALT (CHECK, CURB, STAUNCH, STEM)
– PURPLE: ANIMALS MINUS STARTING LETTER (ANTHER, EASEL, HARK, LAMA)

What’s NYT Connections?
NYT Connections is a fun and challenging word game from the New York Times. You’ll need to find groups of four items that share a common theme, with difficulty levels ranging from easy (green) to very hard (purple). Don’t worry, you can make up to four mistakes and don’t have to solve the hardest group. Just watch out for sneaky word tricks!

Now go forth and conquer today’s NYT Connections puzzle! And if you’re craving more word games, check out our daily Strands hints and answers, Quordle hints and answers, or Marc’s Wordle today page. Happy connecting! 🎉📝