The Download: meet our AI innovators, and what happens when therapists use AI covertly

This is today's edition of the download, our weekday newsletter, which offers a daily dose of what is going on in the world of technology. Meet the AI ​​winners in our 35 innovators under 35 lists for 2025 per year. We choose 35 outstanding people under the age of 35 who use technology to tackle difficult problems in their respective areas. Our AI winners include people who control model development in Silicon Valley's largest technology companies, and academic researchers who develop new techniques to improve AI performance. Take a look at all of our AI innovators and the full list -including our innovator of the year. Like Yichao "Peak" Ji to a global Ki -app -Hitmaker, when Yichao Ji - also became known as "Peak" - in a start video for Manus in March, he did not expect it to be viral. The 32-year-old spoke in fluent English and presented the AI ​​agent, who was built by the Chinese start-up butterfly effect, where he acts as chief scientist. The video was not a detailed production, but something about JIS delivery, and the vision behind the product cut through the noise. The product, which is still available early preview only through invitation codes, spreads into the world within a few days on the Chinese Internet. Within a week after his debut, Manus had attracted a waiting list of around 2 million people. Despite his relative youth, Ji has experience in construction products over a decade that combines technical complexity with real user -friendliness. That brought him credibility - and put him in the foreground of an increasing class of Chinese technologists with global ambitions. Read the whole story. - Caiwei chen! My therapist secretly uses Chatgpt in the imaginary future of Silicon Valley. AI models are so sensitive that we will use them as therapists. They will look after millions for millions who are outrageous of the annoying demands on human consultants, such as the need for graduated degrees, insurance for misconduct and sleep. Something completely different has happened here on Earth. Last week we published a story about people who found that their therapists secretly used chatt during the sessions. In some cases it was not subtle. A therapist accidentally shared his screen during a virtual appointment so that the patient taps into chatt in real time. As a writer of history, Laurie Clarke, it does not emphasize a total pipe dream that AI could be therapeutically useful. But the secret use by therapists of AI models that are not checked for mental health is something completely different. James O’Donnell, our high -ranking AI reporter, had a conversation with Clarke to learn more about what she found. This story originally appeared in the algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, register here. What comes next in technology: the breakthroughs that change some technologies, regardless of whether we are ready or not. Visit us for our Nextedin Live event on September 10, while our editorial team examines the breakthroughs that define this moment, and those on the horizon that demand our attention. From quantum computing to humanoid robotics to AI agents to climate tech, we will examine the innovations that encourage us, the challenges you may bring with you and why you are now worth a reputation. It starts tomorrow at 12.30 p.m. ET and registered here to join us. The must-reads that I combed on the Internet to find them from fun/most important/fascinating/fascinating stories about technology today. 1 The United States gives up its international advance against disinformation, which the Foreign Ministry will no longer work with Europe to combat malicious information that is spread by foreign governments. (Ft $) + It comes to Russia its efforts to insert themselves overseas. (Nyt $) 2 The judges, who monitors the copyright case of Anthropic, is not a happy judge William Alsup that a sealing of $ 1.5 billion may not be in the best interest of the authors. (Bloomberg $) 3 WhatsApp's former security manager, Meta Attaullah Baig sued the company not to protect user data. (Wp $) + he claims that he had uncovered systemic safety errors, but was ignored. (Bloomberg $) + Meta claims that Baig was rejected due to poor performance and not for whistleblowing. (Nyt $) 4 Doge's reigning manager calls on the US government to start setting again after months of widespread surf and resignations. (Fast company $) + How did Doge run chaos in social security. (Prublica) + Doges Technical takeover threatens the safety and stability of our critical data. (With Technology Review) 5 Openai weighs California. It is feared that state supervisory authorities could derail their efforts to convert into a non -profit company. (WSJ $) + Rival Anthropic supports the AI ​​bill by California Governor Gavin Newsom. (Politico) 6 ICE spends millions for facial recognition technology to determine people who suspect that they have attacked officers. (404 Media) + The Supreme Court gave ice to aim for people based on the breed. (VOX) + ICE directors were asked to triple their daily arrests due to undocumented immigrants. (Ny Mag $) 7 AI researchers schools KI to replace them. They record every detail of their working days to help AI record their work. (The information $ $) + people fear that the AI ​​will take on the work of everyone. We have been here before. (With Technology Review) 8 What comes after the smartphone? The rise of AI agents means that we may not stare on glass panels forever. (Nyt $) + What's next for intelligent glasses. (With Technology Review) 9 The obsession of social media with the "locking" must die hectic culture and maximize productivity by all costs, the goals of the game are. (Insider $) 10 As it is to get a massage from a robot, while it may not be completely relaxing, it is relatively cheap. (The guard) + will we ever trust robots? (With Technology Review) quote of the day "It was hell on earth". - Duncan Okindo, who has been enslaved in a cyberscam site in Myanmar and was beaten for missing goals, tells The Guardian about his shattering experience. Another thing that AI means the end of the internet search, as we knew it. We all know what it means to google something colloquially. You pack a few words in a search field and in return receive a list of blues to the most relevant results. Basically, it only gets information that is already on the Internet and shows it in a structured way. But everything that is to be won. We are at a new turning point. The biggest change in the way search engines have given us information since the 1990s is currently taking place thanks to generative AI. Not everyone is happy about the change. Publishers are completely freaked out. And people are also concerned about what these new LLM provider results will mean for our basic common reality. Read the whole story. - Mat Honan We can still have nice things to get comfort, fun and distraction to lighten your day. (Have any ideas? Write me a line or Skeet -Sie on me.) + Stephen King's list of favorite films does not contain much horror. + Switch into a breathtaking live stream on earth and beamed live from the international space station. + The thumbnails of rodents are far more important than I have recognized for + brand, our words, the actor Wagner Moura will be the next big thing.