Le forze armate di Mosca sono tornate a colpire la capitale Kiev e la seconda città più grande del Paese, Kharkiv, oltre a decine di altre località, anche situate centinaia di chilometri dalla linea del fronte.

Paris is rolling out a confrontational online strategy to counter disinformation from the United States and beyond.

Tradotto… Hanno arrestato della gente, hanno guardato le loro chat su Signal, hanno trovato altri che segnalavano la presenza di milizie e adesso li vogliono perseguire per averli “messi in pericolo”. Dopotutto due agenti federali sono morti uccisi a sangue freddo dai terroris … ops. **Chissà perche su Signal e non su messenger, whatsapp o altro… chissà perchè…** Ha fatto bene Meredith Whittaker, presidente della Signal Foundation, ad essersi trasferita in Europa a vivere, anche se non credo sia a rischio di chiusura. Altrimenti come fa il ministro della guerra a vantarsi con gli amici degli attacchi militari che stanno per fare? BTW, ieri guardavo online streaming live di Minneapolis. Ho notato un pattern che e’ successo diverse volte: la linea dei miliziani dell’ICE (“patrioti”) indietreggia, centinaia di manifestanti (“terroristi domestici”) vanno un po’ avanti; ICE spara lacrimogeni. Se qualcuno dei manifestanti dà un calcio ad una lattina di lacrimogeni quelli dell’ICE partono in gruppo di corsa lo blindano a terra, lo ammanettano e lo portano via. Immagino per “impeding” (ostruzione alla giustizia). Stante che guadagnano soldi per ogni arresto che fanno, questo pattern di comportamento si può ben chiamare”business development”. > FBI Director Kash Patel said Monday that he had opened an investigation into the Signal group text chats that Minnesota residents are using to share information about federal immigration agents’ movements, launching a new front in the Trump administration’s conflict there with potential free speech implications.Patel said in an interview with conservative podcaster Benny Johnson that he wanted to know whether any Minnesota residents had put federal agents “in harm’s way” with activities such as sharing agents’ license plate numbers and locations. Source: _FBI investigating MN Signal groups tracking ICE, Patel says_ If you like this post, please consider sharing it.

Che cosa c’è di vero e cosa no nel video in cui lo storico annuncia il suo “no” al referendum costituzionale

Why the human penis is unusually large compared to other primates is a long-standing evolutionary question. This study shows that genital size is a cue used during male-male assessment in humans, sugg...

The service uses large language models and references patients medical records to answer questions, manage medications and book appointments.

Aspettative sbagliate da chi, conoscendo la tecnologia, crede di conoscere il dominio applicativo meglio degli esperti aka la sindrome dell’informatico/consulente “adesso ti spiego io come si fa il tuo lavoro” come è vera la barzelletta sui consulenti… mi diceva il capo della ricerca di una grande azienda farmaceutica che sperano di riuscire a tagliare di circa un anno i tempi. (da una dozzina d’anni a una decina o poco più) > A year ago, Demis Hassabis, founder of Google’s AI drug discovery startup Isomorphic Labs, predicted at Davos that he would have an AI-designed drug in trials by the end of 2025. > > Today, neither Isomorphic nor any of its competitors has reached large-scale clinical trials yet, despite the more than $17 billion of capital invested in the endeavor since 2019. Isomorphic alone raised $579 million last year in a round led by Thrive Capital. > > So, why is it taking so long? Why haven’t the breakthroughs in generative models and robotics we’ve seen over the last couple of years translated into the biomedical sciences? > > … > > But identifying a drug candidate is only the first step in an arduous process. Around 85% of drug candidates fail at the clinical trial stage, by far the longest and most expensive part of the drug development process. … All in, it typically takes between 10 and 15 years for a drug to go from the lab to FDA approval. > > … > > “Most of the failures are because we don’t understand biology,” said Pat Walters, chief scientist of OpenADMET, an open-source project for AI drug discovery. “Is AI going to fix that? Maybe it’s going to help us along the way, but I don’t see it as being a panacea.” > > Still, many new investors have been drawn in by lofty promises of a biotechnology revolution. > > “Investor expectations might have been too large,” said Peter Hebert, managing partner at biotech investor Lux Capital, who said he has seen some of his software-focused peers retreat back to more familiar pastures. Source: _Why AI drug discovery isn’t the layup VCs expected – PitchBook_ If you like this post, please consider sharing it.