Why an AI expert quits Google and regrets his work

We hope you had a good day and now we want to give you some exciting information to take with you into the evening. t3n Daily is also available as a podcast and as a newsletter. Here are the topics of the day.

If anyone knows the potential of AI, it’s Geoffrey Hinton: the deep learning pioneer developed some of the most important techniques in modern AI. But after ten years of research work for Google, he is now leaving the company – and warns of what he helped create. He has “new fears” about the technology he helped develop, he told the New York Times.

In order to be able to speak openly about it, he must be free, he stressed, adding that part of him now “regrets” his life’s work. Hinton had long raised ethical issues in dealing with AI. He was particularly concerned about their use for military purposes. Hinton received the 2018 Turing Award, the equivalent of the Nobel Prize in Computer Science.

Is there water on Mars? This assumption has been around for a long time, but has not yet been proven. The Chinese Mars rover Zhurong, which set off to explore the surface of Mars in July 2020, is now the closest thing to proof. Accordingly, an investigation provided evidence that there could be water at low latitudes. The research team responsible classified the findings obtained as “evidence” with cautious optimism.

However, the researchers failed to provide evidence in the form of detectable ice formed from water. And many questions remain unanswered for the future: “The modern hydroclimatic conditions on Mars are still puzzling, since the explanations for the formation of the observed landforms are not clear,” says the study.

The New York transport company will no longer use Twitter alerts. This was announced by the operating company Metropolitan Transportation Authority (MTA) and justified the decision by saying that the platform was no longer “reliable”. Ten days ago, the Twitter API was suspended and the MTA’s real-time alerts were down.

Another reason might be the high price the company charged Elon Musk for using the API. According to Bloomberg, Twitter’s demands totaled $50,000 a month. The MTA needs to be particularly careful on the money, running a deficit of around $600 million that could grow to $3 billion when federal pandemic aid runs out.

The United States is threatened with a government default as early as June 1, 2023: This worrying news came from none other than US Treasury Secretary Janet Yellen. However, the available reserves could only be used up a few weeks later, she explained.

One thing is clear: the consequences of a US default would be serious. Yellen spoke of the fact that the default could plunge the global economy into a crisis. ECB President Christine Lagarde described such a scenario as a “big, big catastrophe”. If economist Volker Brühl has his way, a stock market crash would follow. Share price drops of ten percent are possible, he says. The dollar exchange rate could also collapse. Ultimately, however, it is unlikely that the USA will allow a default.

Sit down, buckle up: Elon Musk’s lawyers claim that his controversial statements about the safety of the Full Selfdriving assistance system could be deepfakes. They wanted to prevent him from having to testify in court. Musk stated in 2016: “A Model S and Model X can drive autonomously with greater safety than a human at this point in time. Right now.” Autonomy on freeways is “really easy” even with obstacles. He revised these statements years later.

However, the family of a man who died in an accident sued – and has the judge on her side when it comes to the Musk lawyers’ deepfake attempt. The judge ordered a three-hour questioning session and said Tesla’s counter-arguments were “deeply disturbing.”

That’s it for today’s t3n daily. You can find much more about all aspects of digital life, working life and the future around the clock at t3n.de.