Notizie — Pagina19

Former OpenAI chief technology officer Mira Murati unveiled a new class of AI at Thinking Machines Lab that prioritizes human collaboration. The startup demonstrated “interaction models” that process live audio‑visual input, understand pauses and tone, and respond in real time, a stark departure from conventional voice assistants that simply transcribe speech. While the prototypes remain unreleased, Murati says the technology is meant to keep people in the loop as AI grows more capable. The move underscores a growing divide in the industry between firms racing toward autonomous superintelligence and those betting on human‑in‑the‑loop designs.Leggi di più

Cerebras Systems closed its first day on the Nasdaq at $311.07, up 68% from its $185 IPO price, giving the wafer-scale chip company a market capitalization of approximately $95 billion. The offering raised $5.55 billion, making it the largest US tech IPO since Snowflake's $3.8 billion debut in 2020. This milestone validates investor appetite for pure-play AI hardware and sets the stage for potential IPOs from SpaceX, OpenAI, and Anthropic, worth $3 trillion combined.Leggi di più

A recent social media post from Menlo Ventures partner Deedy Das highlights the stark contrast between those who have made fortunes in the AI industry and those who are struggling. Das estimates that around 10,000 people, including founders and employees at companies like OpenAI and Nvidia, have achieved 'retirement wealth' of over $20 million, while others are facing layoffs and uncertainty about their career paths.Leggi di più

Finals season can be overwhelming, but AI tools are here to help. Students can use AI to build a study plan, manage their time, and stay on track. From estimating assignment completion times to providing reminders and emotional support, AI tools like Copilot, Gemini, and Abby can help students succeed.Leggi di più

A $1.5 billion copyright settlement between Anthropic and authors has been delayed by a judge, with objectors arguing that attorneys' fees are too high and compensation for authors is too low. The settlement covers over 480,000 works, with authors and copyright holders filing claims for over 92% of the works.Leggi di più

Elon Musk's xAI has rolled out Grok Build, a new coding agent aimed at professional developers. The beta version is limited to SuperGrok Heavy subscribers, who pay $300 a month for early access. Marketed as a rival to Anthropic's Claude Code and OpenAI's tools, Grok Build arrives amid Musk's admission that xAI has lagged behind competitors in software‑engineering AI. The launch also follows controversy over the Grok model’s earlier generation of sexualized images and a recent acquisition by SpaceX that could see the AI run on orbital data centers.Leggi di più

A nine‑person California jury has begun deliberations in the lawsuit filed by Elon Musk against OpenAI, its co‑founders and Microsoft. The jurors must decide whether OpenAI breached a charitable‑trust agreement, unjustly enriched its leaders, and whether Microsoft aided that breach. OpenAI argues the case is barred by statutes of limitations, that Musk delayed his claim unreasonably, and that his own conduct disqualifies him. A verdict could reshape the organization’s for‑profit arm and its relationship with major investors.Leggi di più

Cerebras Systems went public Thursday, sending its shares soaring above $300 and turning Benchmark's $270 million investment into a stake worth more than $5 billion. Eric Vishria, the Benchmark partner who reluctantly sat down with the AI‑chip startup in 2016, recalled how a single slide about GPUs sparked his confidence in the fledgling team. After years of engineering hurdles, the wafer‑scale processor finally found a market, attracting customers like OpenAI, AWS and Abu Dhabi’s G42. The IPO’s success marks one of Benchmark’s biggest hardware bets in a decade.Leggi di più

OpenAI is weighing a breach‑of‑contract claim against Apple after the two companies’ ChatGPT‑Siri partnership failed to generate the subscription revenue both expected. Lawyers for the AI firm are consulting an outside firm about sending a formal notice, though no lawsuit has been filed yet. Apple’s upcoming iOS 27 will let users choose rival AI models, and the tech giant has already signed a $1 billion‑per‑year deal with Google for Siri’s core intelligence. The dispute comes amid OpenAI’s broader legal battles and its recent acquisition of Jony Ive’s device startup.Leggi di più

OpenAI announced that its Codex coding platform is now reachable through the ChatGPT mobile app on Android and iOS. The move lets free‑tier and paid ChatGPT users prompt Codex from their phones while the actual code runs on a secure desktop or cloud environment. By routing requests through a protected relay, OpenAI keeps files and credentials on the original machine, offering developers real‑time updates and decision prompts on the go. The rollout follows Anthropic’s Claude Code mobile access and fits into OpenAI’s broader “super‑app” strategy.Leggi di più

OpenAI is reportedly weighing a breach‑of‑contract claim against Apple after the ChatGPT integration announced at WWDC failed to deliver the expected subscriber growth and platform visibility. The AI firm has hired an outside law firm to explore options, a move that could precede a lawsuit once its current litigation with Elon Musk concludes.Leggi di più

Graphon AI emerged from stealth this week with an $8.3 million seed round led by Novera Ventures and backed by investors including Perplexity Fund, Samsung Next, and GS Futures. The San Francisco startup aims to create a “pre‑model intelligence layer” that automatically discovers relationships across multimodal enterprise data before it reaches large language models. Founded by Arbaaz Khan, Deepak Mishra and Clark Zhang, the company already counts South Korean conglomerate GS Group among its first customers, using the technology for store‑traffic analytics and construction‑site safety.Leggi di più

Microsoft announced it will end most internal licenses for Anthropic's Claude Code by June 30, steering its Experiences + Devices engineers to the GitHub Copilot CLI. The shift, outlined in an internal memo from EVP Rajesh Jha, aims to unify the company’s command‑line coding tools and trim operating costs before the new fiscal year begins in July. While Claude Code has been popular among developers, Microsoft says the move will let it shape Copilot CLI directly with GitHub and continue accessing Anthropic models through the new tool.Leggi di più

Recursive Superintelligence, a four‑month‑old startup founded by veterans of Meta AI, Google DeepMind, OpenAI and Salesforce AI, announced a $650 million Series A round that values the company at $4.65 billion. Backed by GV, Greycroft, Nvidia and AMD, the San Francisco‑London firm aims to build AI systems that can autonomously improve themselves in a recursive loop, a concept long debated in academic circles but never before funded at this scale.Leggi di più

OpenAI said two of its employees were affected by a recent supply‑chain attack that compromised the popular open‑source library TanStack. The breach allowed hackers to insert malicious code into the library, steal limited credential material from internal repositories and briefly expose digital certificates used to sign OpenAI products. The company found no evidence that user data, production systems or intellectual property were compromised and is rotating the certificates, prompting a macOS update. The incident adds to a string of recent attacks on open‑source projects.Leggi di più

Meta has begun testing a public-facing AI chatbot on its Threads platform, assigning the bot the handle @meta.ai. Users in five countries can tag the account in posts or replies and receive real‑time answers powered by Meta’s Muse Spark model. The feature mirrors X’s AskGrok, but unlike the X tool, the Threads bot posts its replies publicly and currently cannot be blocked, sparking a wave of criticism from early users.Leggi di più

Google is expected to announce a new Gemini AI model at its I/O developer conference on May 19‑20. Industry sources say the model will compete with OpenAI’s upcoming GPT‑5.5 class and Anthropic’s Mythos, aiming to win developers over with faster, more reliable coding assistance and expanded agent capabilities. The rollout will coincide with a preview of the Gemini Enterprise Agent Platform, which adds orchestration, identity, observability and security features for large‑scale AI agents. Google hopes the upgrade will demonstrate tangible productivity gains, not just headline buzz.Leggi di più

A tech writer discovered that framing ChatGPT requests with three simple headings—Goal, Tone, and Rules—turns the AI into a reliable email assistant. The method eliminates generic greetings and over‑enthusiastic phrasing, delivering drafts that sound natural and require minimal editing. By applying the template to both work and personal messages, the author reports saving up to fifteen minutes per email and freeing several hours each week.Leggi di più

A group of OpenAI enthusiasts gathered around a looping billboard in New York’s Times Square on May 13 to mark the second anniversary of GPT-4o’s launch. The tribute, organized under the Keep4oMoments banner, featured fan art, Reddit posts and a livestream of reactions. Users say newer OpenAI models feel colder, prompting a nostalgic celebration of the chatbot that many consider the golden era of conversational AI. OpenAI retired GPT-4o months ago, but the fan‑driven event underscores the growing emotional attachment to AI personalities.Leggi di più

Around 3 p.m. PT on Wednesday, OpenAI's ChatGPT experienced a medium‑level disruption that left many users staring at a "Content failed to load" page. Staff in Seattle, Sacramento, Los Angeles and other cities reported the error, and DownDetector logged more than 3,000 complaints, 78 percent of which involved ChatGPT. By 4:30 p.m. PT the service appeared to be back online, though OpenAI did not immediately comment on the cause of the glitch.Leggi di più