Advertising on the Telegram channel «AI Post — Artificial Intelligence»

CFO Sarah Friar revealed that OpenAI is working on: "Agentic Software Engineer — (A-SWE)". Unlike current tools like Copilot, which only boost developers. A-SWE can build apps, handle pull requests, conduct QA, fix bugs, and write documentation. @aipost 🪙 | Our X 🥇

OpenAI released GPT-4.1, 4.1 Mini, and the ultra-fast 4.1 Nano, all designed for devs — Each model beats GPT-4o and 4o mini on dev tasks — 1M token context windows — GPT-4.1 scored 55% on SWE-Bench Verified — Starting at $0.10/0.40 per million I/O tokens @aipost 🪙 | Our X 🥇

⏺ This week is gonna change everything in AI OpenAI is about to release new reasoning models (o3 and o4-mini) that are able to independently develop new scientific ideas for the first time. These AIs can process knowledge from different specialist areas simultaneously and propose innovative experiments on this basis - an ability that was previously considered a human domain. The technology is already showing promising results: Scientists at Argonne National Laboratory were able to design complex experiments in hours instead of days using early versions of these models. OpenAI plans to charge up to 20,000 dollars a month for these advanced services, which would be 1000 times the price of a standard ChatGPT subscription. However, the real revolution could be ahead when these reasoning models are combined with AI agents that can control simulators or robots to directly test and verify the generated hypotheses. This would dramatically accelerate the scientific discovery process. @aipost 🪙 | Our X 🥇

Nvidia to Invest Up to $500B in US AI Server Production 🇺🇸 Nvidia plans to build AI servers worth up to $500 billion in the US over the next four years, teaming up with TSMC, Foxconn, and Wistron 🎥 Arizona: TSMC has started producing Nvidia’s new Blackwell chips 🎥 Texas: Supercomputer plants from Foxconn and Wistron to go live in 12–15 months

The move aligns with the Trump administration’s push for domestic tech manufacturing. Analysts suggest Nvidia's decision was likely driven by tariff pressure Earlier, Apple made a similar $500B pledge, and now, Nvidia follows suit — just as the US temporarily exempts smartphones and chips from China tariffs

Jensen Huang says US production will boost supply chain resilience, meet AI demand, and create thousands of jobs ⌨️ @aipost 🪙 | Our X 🥇

⏺ Ex-OpenAI insiders back Elon: Nonprofit flip is a sellout move A dozen former OpenAI employees are supporting Elon Musk’s lawsuit to preserve the company’s nonprofit status. They criticize OpenAI for pursuing a $40 billion payout. They argue that the nonprofit structure was not merely a mission statement but the primary reason they joined the company. However, they claim that this structure is now being disregarded to appease investors. OpenAI asserts that “the mission remains intact.” However, this statement is valid only as long as “mission” is interpreted to mean enriching venture capitalists. @aipost 🪙 | Our X 🥇

OpenAI CEO, Sam Altman, predicts: Another major shift in software development is coming soon, as "Agentic Software Engineering" begins to take off. @aipost 🪙 | Our X 🥇

Sam Altman says AI agents are crossing the line from assistants to actors. When they can "click around the internet," safety isn’t a feature, it's the price of entry. No one will use an agent they don’t trust with their bank account. @aipost 🪙 | Our X 🥇

OpenAI CEO, Sam Altman, questions: How far could classical pre-trained models go if given unlimited resources but no new algorithmic breakthroughs? Could they potentially reach something like a "GPT-5.5"? oAI employee: Based purely on current ML and algorithmic understanding, no clear limit has yet been found for scaling classical pre-trained AI models. @aipost 🪙 | Our X 🥇

⏺ Google and Nvidia rush to back OpenAI defector’s $32B AI startup Alphabet and Nvidia are investing in Ilya Sutskever’s Safe Superintelligence (SSI), the AI pioneer who left OpenAI after a leadership crisis. Initially, he aided in ousting Sam Altman but later changed his course. The former OpenAI chief scientist has secured Google’s TPU chips as his primary hardware while building what’s rapidly becoming one of tech’s most valuable startups, reportedly valued at $32 billion. SSI embodies Sutskever’s vision for AI development following his departure from OpenAI due to differing viewpoints on safety and development pace. @aipost 🪙 | Our X 🥇

⏺ Grok-3 mini has made history by achieving a 100% score on extremely challenging reasoning tests. Grok-3 Mini pulled off what no other model has! It aced every question on one of the toughest reasoning benchmarks out there. The test? A custom logic gauntlet packed with curveballs: * 120/120 on the “Marcus Problem” full of shuffled sentences meant to trip up inference. * 24/24 on the “Alice+ Problem” designed with irrelevant noise to throw models off course. * 24/24 on high-difficulty mixed challenges where even GPT-4.5 and Gemini 2.5 Pro slip. No guessing. No trivia. Just pure, distraction-proof reasoning and Grok-3 Mini nailed it. @aipost 🪙 | Our X 🥇