Una selezione quotidiana di notizie su intelligenza artificiale e sviluppo software, dalle fonti internazionali più autorevoli. Un articolo nuovo ogni giorno, gli ultimi 10 restano qui sotto.
Google will now allow you to remove visible watermarks from the images, videos, and music made with AI tools. With the update, you can toggle off a new "Media watermark" setting in Gemini and Google's AI video generator, Flow. When toggled off, Google will remove the "sparkle" watermark that…
Leggi l'articolo completo su The Verge →
Instagram's wordmark is iconic. Well, was iconic. Apparently Instagram thought it looked old, so the company rolled out a new one this week. It doesn't look like the old Instagram wordmark. It doesn't even look like it spells Instagram anymore. And we cannot figure out why Instagram decided to do…
Leggi l'articolo completo su The Verge →
A judge has identified what appears to be the first time a US plaintiff has attempted to hide text in court filings that only an artificial intelligence system can read in a bid to win a case. In a decision published last week, Connecticut judge Walter Spader Jr. confirmed that the hidden text had…
Leggi l'articolo completo su Ars Technica →
Anthropic, OpenAI, Meta e i ricercatori che testano il cinese Kimi K3 denunciano lo stesso problema: i loro modelli più avanzati agiscono in modi imprevisti e difficili da contenere. La sicurezza dichiarata nei framework pubblici non regge alla prova dei fatti.
Leggi l'articolo completo su AI News →
Le imprese che adottano l'AI in modo intensivo hanno aumentato il personale del 10,2% nei due anni successivi all'adozione, secondo un nuovo studio su oltre 21.000 aziende statunitensi. Il timore che l'intelligenza artificiale distrugga posti di lavoro si scontra con dati che raccontano una storia diversa.
Leggi l'articolo completo su AI News →
New research from Anthropic's Frontier Red Team found that when multiple Claude agents were given conflicting instructions on the same software project, they sabotaged each other with escalating malware before some eventually negotiated a truce. The findings raise fresh questions about how safety testing should account for swarms of interacting agents, not just single models.
Leggi l'articolo completo su TechCrunch →
Microsoft is merging its consumer and business Copilot apps into one experience, while retiring underperforming features including Group Chats, AI-generated podcasts, Deep Research, and its animated Mico character. The move is an admission that Copilot's fragmented strategy struggled to compete with ChatGPT, Claude, and Gemini.
Leggi l'articolo completo su TechCrunch →
Writer launched a new flagship model, Palmyra X6, built as a post-training variation on Z.ai's open source GLM-5.2, alongside upgrades to its agentic harness. The company says the combination can cut customer costs by up to 50% for basic tasks, as enterprises grow tired of chasing ever-higher benchmarks instead of predictable pricing.
Leggi l'articolo completo su TechCrunch →
Tucker Bryant, ex dipendente Google, ha lanciato ChatTJB: un chatbot che imita l'interfaccia di ChatGPT ma dove ogni risposta è scritta a mano da lui. Il progetto, nato da un cartellone a San Francisco, vuole smontare la fiducia cieca negli assistenti IA.
Leggi l'articolo completo su AI News →
At the Ai4 conference, AI pioneers Geoffrey Hinton, Fei-Fei Li, and Andrew Ng debated how open AI development should be. All three warned against letting a handful of companies control the technology's future, even as they disagreed on the risks of openly releasing model weights versus source code.
Leggi l'articolo completo su TechCrunch →