Today: 7 February 2026
Subscribe
Browse Category

Artificial Intelligence News 17 July 2025 - 24 July 2025

Cloud Titans Unleash AI Alliances and Mega-Deals – July 23–24, 2025 Cloud News Roundup

Cloud Titans Unleash AI Alliances and Mega-Deals – July 23–24, 2025 Cloud News Roundup

AWS announced an additional $100 million investment into its Generative AI Innovation Center to expand AI engineering capabilities and partner programs. AWS launched the Generative AI Partner Innovation Alliance to streamline AI projects from experimentation to production via a global network of system integrators. A built-in blue/green deployment feature for Amazon ECS was rolled out to enable seamless switching between old and new container versions. A 5-year strategic collaboration between AWS, NatWest Group and Accenture will modernize NatWest’s data, analytics, and AI capabilities for 20 million customers. In Thailand, Microsoft Azure is aiding the government’s legal system overhaul with Azure
24 July 2025
·
Python’s Game-Changer, AI’s Big Letdown & Ada’s Surprise Comeback – Programming News Roundup (July 23-24, 2025)

Python’s Game-Changer, AI’s Big Letdown & Ada’s Surprise Comeback – Programming News Roundup (July 23-24, 2025)

Python 3.14 RC1 introduces an opt-in no-GIL build and an experimental JIT compiler, plus a smarter Windows installer, with single-thread performance in free-threaded mode about 5–10% slower. Java JDK 25 enters Rampdown Phase Two with no new JEPs ahead of its September release. Spring Framework 7.0 reaches its 7th milestone, aligning with JDK 21+ and Jakarta EE 10. .NET 10 Preview 6 debuts with runtime JIT optimizations and JSON serialization improvements, but no new C# or F# features in this preview. TypeScript 5.9 enters beta, and TypeScript 7.0 will introduce a Go-written native compiler for 10x performance, with the first
24 July 2025
AI Breakthroughs, Bold Plans & Backlash: Inside the 48‑Hour Global AI Frenzy (July 23–24, 2025)

AI Breakthroughs, Bold Plans & Backlash: Inside the 48‑Hour Global AI Frenzy (July 23–24, 2025)

On July 23, 2025, President Donald Trump unveiled the U.S. AI Action Plan to pursue “global dominance” by promoting open-source AI models worldwide and slashing regulatory barriers to innovation. The plan includes tracking advanced AI chips to prevent exports to “countries of concern” such as China and links export controls to national security. OpenAI and Oracle announced Project Stargate, adding 4.5 gigawatts of new data-center capacity—enough to run over 2 million AI chips—as part of a plan valued at up to $500 billion to build 10 GW of AI supercomputing across the United States, with OpenAI reporting over 5 GW
24 July 2025
·
Battle of AI Titans: ChatGPT vs Google Gemini – The Ultimate 2025 Showdown

Battle of AI Titans: ChatGPT vs Google Gemini – The Ultimate 2025 Showdown

ChatGPT launched in late 2022 and quickly gained hundreds of millions of users, while Google released Gemini (then Bard) in 2023 as its competing AI assistant. ChatGPT runs on the GPT-4 model and is renowned for strong text reasoning and coding, whereas Gemini is natively multimodal (text, images, audio, video) and can use live Google Search and even execute code in Gemini Pro. Gemini Pro can handle context windows up to 2 million tokens, far larger than ChatGPT’s 32,000-token extended window, enabling whole-project analysis. In February 2025, Gemini Advanced gained the ability to recall past chats to personalize responses, while
23 July 2025
···
AI’s 48-Hour Frenzy: Breakthroughs, Billion-Dollar Bets & Bold Policies (July 22–23, 2025)

AI’s 48-Hour Frenzy: Breakthroughs, Billion-Dollar Bets & Bold Policies (July 22–23, 2025)

Google and OpenAI won gold medals at the International Mathematical Olympiad by solving 5 of 6 problems with new general-purpose reasoning based on natural language understanding. OpenAI disclosed its experimental model scaled up test-time compute to think longer and in parallel, a very expensive undertaking. Google DeepMind’s Gemini Deep Think, unveiled in May, solved IMO problems within the 4.5-hour limit entirely in natural language. MIT CSAIL and Facebook AI researchers presented ICML 2025 work showing image generation with encoder-only tokens, enabling inpainting without a generator and morphing a red panda into a tiger. Alibaba’s Qwen-3 Coder, open-sourced for software development,
23 July 2025
··
AI Accelerator Chip Gold Rush: Inside the Global Race for a $300B Market by 2030

AI Accelerator Chip Gold Rush: Inside the Global Race for a $300B Market by 2030

The AI accelerator chips market is projected to exceed $360 billion by 2032, up from $28.5 billion in 2024, with about 37.4% CAGR. In 2024, GPUs accounted for about 35% of AI accelerator revenue, while custom AI ASICs were the fastest-growing segment at 43%+ CAGR. Cloud data centers dominate AI chip demand at roughly 75%, while edge AI chips are growing at more than 40% annually. North America held about 44% of the AI accelerators market in 2024, and NVIDIA commanded over 80% of data-center AI accelerator hardware. Hyperscalers like Google with TPU and AWS with Inferentia/Trainium have built in-house
21 July 2025
····
Google AI Studio Unleashed: Inside Google’s Gemini-Powered AI Playground Taking on ChatGPT

Google AI Studio Unleashed: Inside Google’s Gemini-Powered AI Playground Taking on ChatGPT

Google AI Studio is a browser-based cloud platform that gives developers direct access to Google DeepMind’s Gemini multimodal models for prototyping and building AI-powered applications. Gemini 1.5 Flash and Gemini 1.5 Pro were in public preview on AI Studio by late 2024, offering up to a 1,000,000-token context window, the largest context length among commercial AI at the time. At Google I/O 2024, Gemini 1.5 Pro extended context to 2,000,000 tokens via a waitlist, doubling single-prompt capacity. In early 2025, Gemini 2.0 was released with Flash and Flash-Lite variants for low-latency production use and a 2.0 Pro Experimental focused on
21 July 2025
·
Major AI Conferences, Exhibitions, and Awards in 2025–2026

Major AI Conferences, Exhibitions, and Awards in 2025–2026

AAAI 2025 was held February 27–March 2, 2025 in Philadelphia, USA, and AAAI 2026 is scheduled for January 20–27, 2026 in Singapore. NeurIPS 2025 will be held December 10–15, 2025 in San Diego, USA, attracting over 13,000 participants. ICML 2025 takes place July 13–19, 2025 in Vancouver, Canada, with around 6,000 participants. ICLR 2025 is scheduled for April 24–28, 2025 in Singapore, marking the conference’s first return to Asia post-pandemic. IJCAI 2025 runs August 16–22, 2025 in Montréal, Canada, and IJCAI-ECAI 2026 will be a joint conference in Bremen, Germany, August 8–14, 2026, with about 2,500 attendees. CVPR 2025 takes
21 July 2025
···
Popcorn-Serving Robots, Brain Chips & Billions in AI: Inside Elon Musk’s July 2025 Robotics Revolution

Popcorn-Serving Robots, Brain Chips & Billions in AI: Inside Elon Musk’s July 2025 Robotics Revolution

In June, Milan Kovac, the senior VP heading Tesla’s Optimus program, left the company, leaving the project reportedly “in disarray”. Tesla paused parts orders for about two months to implement a major redesign after finding the current Optimus prototype isn’t very useful in factory tasks. Elon Musk had initially targeted 5,000–10,000 Optimus robots in 2025, but that goal is now delayed. Ashok Elluswamy, Tesla’s Autopilot chief, was appointed to take over the Optimus robot program. Key hardware issues include overheating actuators in joint motors, weak payload in the robot’s dexterous hands, premature wear in transmission components, and limited battery life.
21 July 2025
····
Meta’s AI Revolution Unleashed: July 2025 Breakthroughs, Billion‑Dollar Bets & Backlash

Meta’s AI Revolution Unleashed: July 2025 Breakthroughs, Billion‑Dollar Bets & Backlash

In July 2025, Meta AI expanded to include generative video editing powered by the MovieGen model, with one-click effects to change outfits, backgrounds, and lighting. WhatsApp introduced an AI concierge for business chats that handles customer inquiries and post-purchase recommendations, announced at Conversations on July 1, with testing in select markets including Mexico and plans for cloud-based voice and video calling. The Ray-Ban smart glasses’ companion app was rebranded as the Meta AI app, enabling hands-free AI queries and moment recording, following the June debut of Oakley Meta Glasses with built-in AI. In mid-July 2025, Meta and AWS announced a
21 July 2025
··
AI in Overdrive: Breakthroughs, Billion-Dollar Bets, and Bold Policies – News Roundup (July 20–21, 2025)

AI in Overdrive: Breakthroughs, Billion-Dollar Bets, and Bold Policies – News Roundup (July 20–21, 2025)

OpenAI’s experimental large language model earned a gold medal at the 2025 International Math Olympiad, solving 5 of 6 problems for 35/42 points, with OpenAI’s Alexander Wei hinting that GPT-5 will be a system of multiple specialized models orchestrated by a routing algorithm. ChatGPT introduced an autonomous ‘Agent’ mode on July 17, enabling web browsing, plugins, and online purchases, though it remains unavailable in the EU due to regulatory uncertainty. AWS unveiled ‘AgentCore’ for building autonomous AI agents, including a sandboxed code interpreter and integrated web browser, and pledged $100 million to fund ‘agentic AI’ startups via its Generative AI
21 July 2025
·
The Future Is Here: AI’s Most Shocking Developments on July 20, 2025

The Future Is Here: AI’s Most Shocking Developments on July 20, 2025

OpenAI’s latest experimental reasoning model achieved a gold medal performance at the International Math Olympiad by solving 5 of 6 problems under contest conditions. GPT-5 is reportedly on the horizon as a system of specialized models with a smart router, not a single monolith. Demis Hassabis, CEO of DeepMind, predicted AI could reach human‑level problem‑solving within five years, calling the shift bigger than the Industrial Revolution. NVIDIA and university labs introduced DiffusionRenderer, a tool that can insert CGI elements into real videos, demonstrated with an AI-generated dragon casting lifelike shadows. OpenAI rolled out ChatGPT Agent mode, letting the chatbot browse
20 July 2025
·······
Artificial Intelligence in the Military: How AI Is Reshaping the Future of War

Artificial Intelligence in the Military: How AI Is Reshaping the Future of War

The Pentagon’s Project Maven, now managed by the National Geospatial-Intelligence Agency, uses AI to identify objects in surveillance imagery and speeds target identification timelines from hours to minutes. In 2024, the U.S. tested Scylla AI at Blue Grass Army Depot, which autonomously detected intruders and flagged a guard in seconds, achieving over 96% accuracy distinguishing threats from benign activities. Maven’s user base expanded to tens of thousands of military analysts across all combatant commands. The U.S. Department of Defense updated its Autonomous Weapons Policy with Directive 3000.09 in January 2023, adding ethical safeguards for AI-enabled weapons. Joint All-Domain Command and
20 July 2025
·
AI Weekend Shockwave: Global Breakthroughs, Big Tech Bets & Bold Moves (July 19–20, 2025)

AI Weekend Shockwave: Global Breakthroughs, Big Tech Bets & Bold Moves (July 19–20, 2025)

OpenAI rolled out ChatGPT Agent mode, allowing the chatbot to take actions on a user’s behalf using a built‑in browser and plugins, with access granted to paying subscribers. AWS announced AgentCore, a toolkit to build enterprise AI agents at scale, featuring seven agent services and an AI Agents Marketplace, plus a $100 million fund for agentic AI startups. Meta Platforms formed a new Superintelligence Labs unit and vowed to invest hundreds of billions in AI, hiring Mark Lee and Tom Gunter from Apple and Alexandr Wang of Scale AI as part of an aggressive recruitment drive for AGI, and planning
20 July 2025
Google Pixel 10 Leaks Reveal 5× Zoom Camera, New AI Tricks & Early Launch – Everything We Know

Google Pixel 10 Leaks Reveal 5× Zoom Camera, New AI Tricks & Early Launch – Everything We Know

The Google Pixel 10 launch event is scheduled for August 20, 2025 in New York City, where the Pixel 10 lineup will be unveiled. For the first time on a base Pixel, Pixel 10 adds a telephoto camera with around 5× optical zoom alongside a 50 MP Samsung GN8 main sensor and a 13 MP Sony IMX712 ultrawide. The Pixel 10 base features a 6.3-inch flat OLED display with 120 Hz, while the Pixel 10 Pro uses LTPO 1–120 Hz with brighter panels and the Pro XL sits at 6.8 inches with similar specs. The Tensor G5 chip uses TSMC’s
20 July 2025
·
AI Upheaval: 48 Hours of Breakthroughs, Big Tech Bets & Backlash (July 18–19, 2025)

AI Upheaval: 48 Hours of Breakthroughs, Big Tech Bets & Backlash (July 18–19, 2025)

OpenAI rolled out ChatGPT Agent on July 18–19, 2025, a mode that lets the chatbot act on the user’s behalf using a virtual browser and plugins like Gmail and GitHub to perform multi-step tasks with user permission. AWS unveiled AgentCore at its NY Summit, a toolkit for enterprises to build and deploy custom AI agents at scale, featuring seven core agent services and an AI Agents Marketplace, plus a $100 million fund for agentic AI startups. Meta Platforms formed Superintelligence Labs and vowed to invest hundreds of billions of dollars in AI, while carrying out a mass talent recruitment including
19 July 2025
Understanding How AI Works in 2025: Core Concepts, Breakthroughs, and Societal Impact

Understanding How AI Works in 2025: Core Concepts, Breakthroughs, and Societal Impact

By 2023, the U.S. FDA had approved 223 AI-powered medical devices, up from 6 in 2015. GPT-4, released in 2023, is described as having over 100 trillion parameters and powers ChatGPT-like capabilities. The Transformer architecture was introduced in 2017 with the paper “Attention Is All You Need,” enabling self-attention and the training of very large models. Diffusion models such as DALL-E 2, Stable Diffusion, and Midjourney, introduced in 2022–2023, revolutionized image generation. AlphaFold solved the protein folding problem in 2021, and by 2025 its database provides structures for nearly all catalogued proteins. By 2025, 78% of organizations report using AI
18 July 2025
····
OpenAI Unleashes ChatGPT Agent – The AI Assistant That Can Plan, Shop and Make PowerPoints For You

OpenAI Unleashes ChatGPT Agent – The AI Assistant That Can Plan, Shop and Make PowerPoints For You

OpenAI launched ChatGPT Agent on July 17, 2025, turning ChatGPT into a personal AI assistant capable of autonomously handling multi-step tasks. The agent uses a built-in “virtual computer” to browse the web, fill forms, run code, and produce Excel spreadsheets or PowerPoint presentations on the user’s behalf. It can manage schedules and plans by checking your calendar and finding free times to reserve tables via OpenTable. It can perform online research and compile concise reports or analyses, including trend summaries. It can shop online, compare options with criteria you specify, and place orders with your permission. It can generate editable
17 July 2025
AlphaEvolve: DeepMind’s Gemini-Powered AI That Invents Algorithms and Breaks a 56-Year Record

AlphaEvolve: DeepMind’s Gemini-Powered AI That Invents Algorithms and Breaks a 56-Year Record

In mid-2025, Google DeepMind unveiled AlphaEvolve, an AI coding agent that uses Google’s Gemini LLMs, an evolutionary search process, and automated testing to autonomously design and improve algorithms. AlphaEvolve generates, tests, and evolves programs in an autonomous pipeline, refining entire codebases rather than a single function. It discovered a 4×4 matrix multiplication method over the complex domain using 48 multiplications, beating Strassen’s 49 from 1969 for the first time in 56 years. It devised a data-center scheduling algorithm that recovers about 0.7% of Google’s worldwide compute capacity, with the code in production for over a year. In hardware, it produced
17 July 2025
···
32B AI Model Trained by a Swarm of Volunteer GPUs – Inside INTELLECT-2’s Decentralized Revolution

32B AI Model Trained by a Swarm of Volunteer GPUs – Inside INTELLECT-2’s Decentralized Revolution

In May 2025, Prime Intellect unveiled INTELLECT-2, a 32B-parameter LLM trained on a globally distributed swarm of volunteer GPUs. INTELLECT-2 is the first model of its scale trained via fully asynchronous reinforcement learning across hundreds of heterogeneous, permissionless machines on chakra.dev. PRIME-RL coordinates the training loop by separating experience generation on inference workers from policy updates on trainer nodes, eliminating traditional synchronization bottlenecks. SHARDCAST distributes the 32B model weights via a tree-based, pipelined transfer to hundreds of volunteer GPUs, accelerating weight propagation. TOPLOC provides a probabilistic verification fingerprint using locality-sensitive hashing to detect tampering or incorrect results from untrusted nodes.
17 July 2025
1 … 137 138 139 140 141 … 144

Stock Market Today

Go toTop