Latest Tech News

Stay updated with the latest in technology, AI, cybersecurity, and more

Filtered by: del Clear Filter

Leaker says iPhone 17 lineup getting two new screen changes

Two months from now, Apple is expected to unveil its new flagship iPhone lineup, including iPhone 17, iPhone 17 Air, iPhone 17 Pro, and iPhone 17 Pro Max. And today, a leaker has shared two screen changes that may be coming to various new models. Today in a new post on Weibo, leaker Digital Chat Station has shared a variety of expectations for the upcoming iPhone 17 line. Many of the details have been reported before, but there are two interesting new details that stand out most—both relating

Topics: 17 iphone models new pro

If You Love LEGO, the Star Wars Millennium Falcon Is at a Record Low Price for Prime Day

LEGO sets have become a true phenomenon during events like Prime Day and Black Friday, mainly because these sales are often the only times you’ll see real discounts on the most popular models. This year, Amazon is offering record low prices on LEGO’s two biggest franchises: Star Wars and Harry Potter. Among the top sellers, the LEGO Star Wars Millennium Falcon A New Hope 25th Anniversary model is one of the most demanded sets. It’s inexpensive, enjoyable to construct, readily recognizable and a

People Are Using AI Chatbots to Guide Their Psychedelic Trips

Trey had struggled with alcoholism for 15 years, eventually drinking heavily each night before quitting in December. But staying sober was a struggle for the 36-year-old first responder from Atlanta, who did not wish to use his real name due to professional concerns. Then he discovered Alterd, an AI-powered journaling app that invites users to “explore new dimensions” geared towards psychedelics and cannabis consumers, meditators, and alcohol drinkers. In April, using the app as a tripsitter—a

OpenAI says GPT-5 will unify breakthroughs from different models

OpenAI has again confirmed that it will unify multiple models into one and create GPT-5, which is expected to ship sometime in the summer. ChatGPT currently has too many capable models for different tasks. While the models are powerful, it can be confusing because all models have identical names. But another issue is that OpenAI maintains an "o" lineup for reasoning capabilities, while the 4o and other models have multi-modality. With GPT-5, OpenAI plans to unify the breakthrough in its lineu

I don't think AGI is right around the corner

“Things take longer to happen than you think they will, and then they happen faster than you thought they could.” - Rudiger Dornbusch I’ve had a lot of discussions on my podcast where we haggle out timelines to AGI. Some guests think it’s 20 years away - others 2 years. Here’s where my thoughts stand as of June 2025. Continual learning Sometimes people say that even if all AI progress totally stopped, the systems of today would still be far more economically transformative than the internet.

I extracted the safety filters from Apple Intelligence models

Decrypted Generative Model safety files for Apple Intelligence containing filters Structure decrypted_overrides/ : Contains decrypted overrides for various models. com.apple.*/ : Directory named using the Asset Specifier assosciated with the safety info Info.plist : Contains metadata for the override AssetData/ : Contains the decrypted JSON files : Contains decrypted overrides for various models. get_key_lldb.py : Script to get the encryption key (see usage info below) : Script to get the en

Reinforcement Learning from Human Feedback (RLHF) in Notebooks

Reinforcement Learning from Human Feedback (RLHF) in Notebooks This repository provides a reference implementation for Reinforcement Learning from Human Feedback (RLHF) [Paper] framework presented in the RLHF from scratch, step-by-step, in code YouTube video. Overview of RLHF RLHF is a method for aligning large language models (LLMs), like GPT-3 or GPT-2, to better meet users' intents. It is essentially a reinforcement learning approach, where rather than directly getting the reward or feedba

This Dell 15″ Laptop (i5, 1TB PCIe SSD, 32GB RAM) Is 75% Off, Amazon Is Going Nuts for Prime Day

There is no better time to shop for laptops on Amazon than Prime Day. As the biggest shopping event of the year for the retailer, Prime Day is when Amazon slashes prices on everything – even going as far as selling best-selling items at a loss. This year, the Dell 15-inch 3530 laptop (Intel 10-Core i5-1334U, 1TB PCIe SSD, 32GB DDR4 RAM) is the best example of just how low these prices can go: You can pick up this beast of a machine right now for only $649, a whopping 73% off its regular price o

Yet Another Zip Trick

To delete account insert your password. After that, a e-mail with deletion confirm link will be send. Click on it to confirm account deletion. You already have a account? and to the I have read and agree to the Do you consent to our site using cookies in accordance to our Privacy Policy? You can consent either to all cookies, or accept only selected types. Reject optional Cache only Allow all cookies Advanced cookie settings

Just Ask for Generalization (2021)

Generalizing to what you want may be easier than optimizing directly for what you want. We might even ask for "consciousness". This blog post outlines a key engineering principle I’ve come to believe strongly in for building general AI systems with deep learning. This principle guides my present-day research tastes and day-to-day design choices in building large-scale, general-purpose ML systems. Discoveries around Neural Scaling Laws, unsupervised pretraining on Internet-scale datasets, and o

Optimizing Tool Selection for LLM Workflows with Differentiable Programming

Modern agentic architectures rely heavily on chaining LLM calls. A typical pattern looks like: Use an LLM to decide which tool to invoke Call the tool (e.g. search, calculator, API) Use another LLM call to interpret the result and generate a final response This structure is easy to reason about, simple to prototype, and generalizes well. But it scales poorly. Each LLM call incurs latency, cost, and token overhead. More subtly, it compounds context: every step includes not only the original q

Deltarune is halfway done, but it feels like a full game

is a news editor covering technology, gaming, and more. He joined The Verge in 2019 after nearly two years at Techmeme. It’s been an agonizing wait for more Deltarune. The first chapter of the episodic, parallel story to Toby Fox’s acclaimed RPG Undertale came out nearly seven years ago. But the newly released third and fourth chapters are so good that I highly recommend playing everything that’s available right now — even if there are still three more episodes to go. In Deltarune, you play as

A new, faster DeepSeek R1-0528 variant appears from German lab

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now It’s been a little more than a month since Chinese AI startup DeepSeek, an offshoot of Hong Kong-based High-Flyer Capital Management, released the latest version of its hit open source model DeepSeek, R1-0528. Like its predecessor, DeepSeek-R1 — which rocked the AI and global business communities with how cheaply it was trained and how wel

Prompting LLMs is not engineering

Prompting LLMs is not engineering published in: With the proliferation of AI models and tools, there's a new industry-wide fascination with snake oil remedies called "prompt engineering". As of July 2025 the term is now "context engineering" or "context prompting" or "context manipulation". To put it succinctly, prompt engineering is nothing but an attempt to reverse-engineer a non-deterministic black box for which any of the parameters below are unknown: training set weights constraints o

Apple just released a weirdly interesting coding language model

Apple quietly dropped a new AI model on Hugging Face with an interesting twist. Instead of writing code like traditional LLMs generate text (left to right, top to bottom), it can also write out of order, and improve multiple chunks at once. The result is faster code generation, at a performance that rivals top open-source coding models. Here’s how it works. The nerdy bits Here are some (overly simplified, in the name of efficiency) concepts that are important to understand before we can move

Inside India’s scramble for AI independence

Nonetheless, a small but determined group of Indian builders is starting to shape the country’s AI future. For example, Sarvam AI has created OpenHathi-Hi-v0.1, an open-source Hindi language model that shows the Indian AI field’s growing ability to address the country’s vast linguistic diversity. The model, built on Meta’s Llama 2 architecture, was trained on 40 billion tokens of Hindi and related Indian-language content, making it one of the largest open-source Hindi models available to date.

WASM Agents: AI agents running in the browser

One of the main barriers to a wider adoption of open-source agents is the dependency on extra tools and frameworks that need to be installed before the agents can be run. In this post, we show how to write agents as HTML files, which can just be opened and run in a browser. One of the main barriers to a wider adoption and experimentation with open-source agents is the dependency on extra tools and frameworks that need to be installed before the agents can be run. In this post, we introduce the

xAI prepares Grok 4 Code as it plans to take on Claude and Gemini

xAI is preparing the rollout of Grok 4, which replaces Grok 3 as the new state-of-the-art model. Ahead of the rollout, testers on X have again spotted references to a few new Grok 4 models. One of the models is called "grok-4-prod-mimic," which reportedly excels at enterprise use cases like data extraction, coding, and text summarisation. It also possesses deep domain knowledge in finance, healthcare, law, and science. But "grok-4-prod-mimic" isn't the only new model coming to Grok. Another

Topics: called grok model new xai

Sakana AI’s TreeQuest: Deploy multi-model teams that outperform individual LLMs by 30%

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Japanese AI lab Sakana AI has introduced a new technique that allows multiple large language models (LLMs) to cooperate on a single task, effectively creating a “dream team” of AI agents. The method, called Multi-LLM AB-MCTS, enables models to perform trial-and-error and combine their unique strengths to solve problems that are too complex

Topics: ab ai mcts model models

The End of Moore's Law for AI? Gemini Flash Offers a Warning

For the past few years, the AI industry has operated under its own version of Moore's Law: an unwavering belief that the cost of intelligence would perpetually decrease by orders of magnitude each year. Like clockwork, each new model generation promised to be not only more capable but also cheaper to run. Last week, Google quietly broke that trend. In a move that at first went unnoticed, Google significantly increased the price of its popular Gemini 2.5 Flash model. The input token price double

Dust hits $6M ARR helping enterprises build AI agents that actually do stuff instead of just talking

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Dust, a two-year-old artificial intelligence platform that helps enterprises build AI agents capable of completing entire business workflows, has reached $6 million in annual revenue — a six-fold increase from $1 million just one year ago. The company’s rapid growth signals a shift in enterprise AI adoption from simple chatbots toward sophi

CoreWeave is the first cloud provider to deploy Nvidia's latest AI chips

Nvidia's Blackwell Ultra chips, the company's next-generation graphics processor for artificial intelligence, have been commercially deployed at CoreWeave, the companies announced on Thursday. CoreWeave has received shipments of Dell-built shipments based around Nvidia's GB300 NVL72 AI systems, Dell said on Thursday. It's the first cloud provider to install systems based around Blackwell Ultra. The Blackwell Ultra is Nvidia's latest chip, expected to ship in volume during the rest of the year.

New Mac models through 2026 revealed in leaked Apple identifiers

Apple still has plenty of new product launches coming through the rest of the year, including at least two Macs. But today, a new report has outlined Mac models coming not only in 2025, but even those expected through 2026. Report includes 15 forthcoming Mac models expected through 2026 Marko Zivkovic writes at AppleInsider: Through our own data and information shared with us, AppleInsider has learned of 15 new Mac models currently in development. People familiar with the matter — and common

Topics: apple inch mac models pro

HOLY SMOKES! A new, 200% faster DeepSeek R1-0528 variant appears from German lab TNG Technology Consulting GmbH

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now It’s been a little more than a month since Chinese AI startup DeepSeek, an offshoot of Hong Kong-based High-Flyer Capital Management, released the latest version of its hit open source model DeepSeek, R1-0528. Like its predecessor, DeepSeek-R1 — which rocked the AI and global business communities with how cheaply it was trained and how wel

How to Choose the Right TV: Quantum Dots, HDR, RGB, and More in 2025

TVs can be tricky to buy. Whether you read our exquisite guide to the Best TVs or not, you probably won't find every top model at your local Walmart, Best Buy, or Costco. And when you're browsing retail websites, it's easy to feel overwhelmed by marketing speak like HDR, QLED, or “local dimming.” That's why we've collected a few tips, tricks, and terms to help you shop smarter when buying a new TV. Think of this as a checklist to consider before you buy your next big screen. Be sure to check ou

Topics: 8k models screen tv tvs

Amazon’s Big Price Drop Makes Dell’s Big 27-Inch Monitor a 4th of July Steal, Skip Waiting for Prime Day

There are worse ways to kick off a long holiday weekend than by grabbing a great Amazon tech deal on a big monitor that’s long on great features, especially when it’s a Dell monitor that’s just $140. The Dell S2725HS is a 27-inch beauty with an ultra-thin design, terrific HD resolution, outstanding sound from built-in 5W speakers, and many more features that make it an absolute bargain at just $140. The fast 100Hz refresh rate and premium reengineered sound that comes from this Dell 27-inch mon

I'm Rethinking My Upgrade Based on These 5 iPhone 17 Pro Max Rumors

I've always been on team Regular Size when it comes to iPhone Pro models, even in years when the larger Pro Max models have gained exclusive features, like the 5x telephoto "tetraprism" camera in the iPhone 15 Pro Max (which expanded to both models in the iPhone 16 Pro). Despite the more expansive screen size, the Max has always seemed too bulky to carry around all day. But a host of rumored features and design changes for the upcoming iPhone 17 Pro Max has me weighing the benefits of embracing

Uber Adds Rental Car Delivery and Discounted Airport Rides Before July 4 Weekend Surge

With AAA forecasting that a record-breaking 72 million Americans will travel during the week of July 4, ride-hailing and delivery company Uber is launching various updates to make summer vacations more seamless. Uber is significantly expanding its rental-car delivery service, introducing new airport options for ride-sharing and prescheduled pickups, and adding more flexibility for hotel food deliveries. Read also: Uber's Simplified App for Seniors Is Launching Nationwide Car rental dropoff exp

I scanned all of GitHub's "oops commits" for leaked secrets

TL;DR GitHub Archive logs every public commit, even the ones developers try to delete. Force pushes often cover up mistakes like leaked credentials by rewriting Git history. GitHub keeps these dangling commits, from what we can tell, forever. In the archive, they show up as “zero-commit” PushEvents . I scanned every force push event since 2020 and uncovered secrets worth $25k in bug bounties. Together with Truffle Security, we're open sourcing a new tool to scan your own GitHub organization for

Could Google’s Veo 3 be the start of playable world models?

Demis Hassabis, CEO of Google’s AI research organization DeepMind, appeared to suggest Tuesday evening that Veo 3, Google’s latest video-generating model, could potentially be used for video games. In response to a post on X beseeching Google to “Let me play a video game of my veo 3 videos already,” and asking, “playable world models wen?” Hassabis responded, “now wouldn’t that be something.” On Wednesday morning, Logan Kilpatrick, lead product for Google’s AI Studio and Gemini API, chimed in