Ilya Sutskever isn't done with AI safety yet

AI News


Welcome to TechCrunch's Week in Review, TechCrunch's newsletter rounding up the biggest stories from the week. Want to get it in your inbox every Saturday? Sign up here.

This week, Ilya Sutskever founded a new AI company, Safe Superintelligence Inc. (SSI), just one month after officially leaving OpenAI. Sutskever, along with Jan Reicke, were integral to OpenAI's efforts to improve AI safety as AI systems become “superintelligent.” However, dramatic disagreements with leadership over how to approach AI safety led to both Sutskever and Reicke leaving the company.

In EV news, Fisker has filed for Chapter 11 bankruptcy, ending months of problems with its Ocean SUV, including recalls and dozens of lemon law lawsuits. This is the second time Henrik Fisker's eponymous car company has found itself in bankruptcy; his first effort began in 2007 and filed for bankruptcy protection in 2013.

Change Healthcare confirmed this week that the medical records of a “significant proportion of Americans” were stolen in a ransomware attack in February. The company processes insurance and billing for patients at thousands of hospitals, pharmacies and clinics, giving it access to vast amounts of health information for roughly a third of all Americans.

news

Department of Justice v. Adobe: The U.S. Department of Justice has filed a lawsuit against Adobe, alleging that the company hid cancellation fees and made it difficult to cancel subscriptions. read more

OpenAI Acquires Rockset: As OpenAI continues to invest in its enterprise sales and technology organization, it announced it has acquired Rockset, a developer of tools to power real-time search and data analytics. read more

The button is back: Clicks has released a nostalgic BlackBerry-inspired phone case that adds a keyboard with physical buttons to the bottom of your iPhone, and we got to try it out for ourselves. read more

Where humans and AI coexist: Butterflies is a social network where humans and AI interact with each other through posts, comments, and DMs in order to build more creative relationships with AI. read more

Apple to discontinue Pay Later: Apple's Pay Later feature will be phased out after its launch in late March 2023. Instead, Apple Pay users will be able to access loans through a partnership with third-party app Affirm. read more

Outlook users take note: Researchers have discovered a bug that could allow anyone to spoof Microsoft corporate email accounts, making phishing attacks more likely to appear credible and fool their targets. read more

Confusion hits Google: The AI-powered search startup currently surfaces results of fact-based questions such as weather and time in a location, currency conversions, and answers to simple math questions directly through cards. read more

Runway Launches Gen-3: The company's latest video generation AI model boasts “significant” improvements in speed and greater control over the structure, style and movement of the videos it generates. read more

analysis

What should AI look like? From black holes to colorful blobs, AI can be difficult to represent in user interfaces. Approaches to branding an all-seeing, all-knowing, all-doing intelligence vary, but Devin Coldewey explores how companies have united around the idea that AI avatars should be non-threatening and abstract, yet relatively simple and non-anthropomorphic. read more

Why Fisker Failed: As Fisker files for Chapter 11 bankruptcy protection, many are wondering about the future of the ill-fated EV startup. Sean O'Kane argues that whatever happens to Fisker or its assets, it won't change the fundamental problem: it wasn't prepared to tackle bringing a flawed vehicle to market. read more

Pushing the cultural boundaries of ChatGPT: ChatGPT currently provides overgeneralized answers to specific questions for specific communities, and its training seems Eurocentric and Western biased. As most AI models are not built with people of color in mind, Dominic-Madori Davis and Tage Kene-Okafor report on a Black-owned chatbot and version of ChatGPT specifically for Black and Brown communities to help founders capitalize on OpenAI's cultural mistakes. read more



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *