😮😱

AI’s biggest value isn’t cutting costs. It’s multiplying what a company can do. The real question isn’t “What can we save?” It’s “What can we do now that wasn’t possible before?” View our new blog post: creagia.com/blog/ai-is-not-j…

One common issue with personalization in all LLMs is how distracting memory seems to be for the models. A single question from 2 months ago about some topic can keep coming up as some kind of a deep interest of mine with undue mentions in perpetuity. Some kind of trying too hard.

Quina bona pinta!

Aquesta setmana he parlat d’IA davant de 400 persones a l’AUGA. Hem vist què és, com funciona i com ens afecta. Tema complex, interessant i amb molt debat. I ho vaig tancar així: “El problema no és que les màquines aprenguin a pensar, sinó que les persones deixem de fer-ho.”

Apple Google = Siri amb cervell “extern”. Apple reconeix que construir un model d’IA capaç des de zero és més car i lent que col·laborar amb Google Gemini per fer que Siri sigui realment útil.

Introducing Cowork: Claude Code for the rest of your work. Cowork lets you complete non-technical tasks much like how developers use Claude Code.

Today we’re bringing @Gmail into the Gemini era, making it a personal, proactive inbox assistant to help you manage your life, not just your messages. Explore the new features launching today, many of which are made possible by Gemini 3 🧵⬇️

😮🤯

Too many companies think “we need an app” or “we need a CRM” as if software was the goal. But software is never the end — it’s just the means to solve problems and achieve business goals. 👉 Read more: creagia.com/blog/why-softwar…

Two weeks ago, @brave just demoed how a prompt injection can hijack an AI model: 📝 A hidden instruction in text makes the AI ignore its rules & follow malicious orders. Their new system Comet acts as a “semantic firewall” to stop it. vimeo.com/1111446047

This That Gemini = Magic Gemini now lets you upload multiple pictures and combine them into one image. Just upload your pics and prompt to combine them. You can get specific about what to change, or let Gemini do its thing. Share your creations in the replies 👇

Last week the @WhiteHouse 🇺🇸 created a Chief Design Officer role a National Design Studio. Goal: redesign gov digital services to be clearer, more consistent & user-friendly. A reminder that UX is not “cosmetic” — it’s trust whitehouse.gov/presidential-… #UX #DigitalDesign

Bank of America inverteix 13.000 M$/any en tecnologia. El seu chatbot d’IA Erica ha superat els 3.000 M d’interaccions amb clients i la versió interna és usada pel 90% dels 213.000 empleats. Quan la tecnologia és útil i usable 👉 🚀 wsj.com/articles/the-man-beh… #AI #Tech #Banking

😵‍💫🤯

Ha estat un plaer poder estar avui a l’Stripe Live Madrid 2025 en nom de @creagia Un esdeveniment complert i interessant al voltant de la infraestructura de pagaments a Internet. Ens tornem a veure l’any vinent! @stripe

GPT-4.5 is ready! good news: it is the first model that feels like talking to a thoughtful person to me. i have had several moments where i've sat back in my chair and been astonished at getting actually good advice from an AI. bad news: it is a giant, expensive model. we really wanted to launch it to plus and pro at the same time, but we've been growing a lot and are out of GPUs. we will add tens of thousands of GPUs next week and roll it out to the plus tier then. (hundreds of thousands coming soon, and i'm pretty sure y'all will use every one we can rack up.) this isn't how we want to operate, but it's hard to perfectly predict growth surges that lead to GPU shortages. a heads up: this isn’t a reasoning model and won’t crush benchmarks. it’s a different kind of intelligence and there’s a magic to it i haven’t felt before. really excited for people to try it!

Introducing Willow, our new state-of-the-art quantum computing chip with a breakthrough that can reduce errors exponentially as we scale up using more qubits, cracking a 30-year challenge in the field. In benchmark tests, Willow solved a standard computation in <5 mins that would take a leading supercomputer over 10^25 years, far beyond the age of the universe(!).

Site search is used more than 4 million times a month and is one of the main ways that people find information on @GOVUK. Find out how we've been improving site search, one of our priorities from our strategy for growth, in our latest blog post 👇 insidegovuk.blog.gov.uk/2024…

realtime api (speech-to-speech): openai.com/index/introducing… vision in the fine-tuning api: openai.com/index/introducing… prompt caching (50% discounts and faster processing for recently-seen input tokens): openai.com/index/api-prompt-… model distillation (!!): openai.com/index/api-model-d…