Welcome again to Ztoog’s Week in Review — Ztoog’s publication recapping the week’s greatest information. Want it in your inbox each Saturday? Sign up right here.
This week, Ilya Sutskever launched a brand new AI firm, Safe Superintelligence Inc. (SSI), only one month after formally leaving OpenAI. Sutskever, alongside Jan Leike, was integral to OpenAI’s efforts to enhance AI safety with the rise of “superintelligent” AI methods. Yet each Sutskever and Leike left the corporate after a dramatic falling-out with management over easy methods to method AI safety.
In EV information, Fisker filed for Chapter 11 chapter safety, capping off months of issues with its Ocean SUV that included recollects and dozens of lemon legislation lawsuits. This is the second automobile firm Henrik Fisker named after himself that has wound up in chapter. His first effort began in 2007 and filed for chapter safety in 2013.
Change Healthcare this week confirmed {that a} February ransomware assault resulted within the theft of medical information affecting a “substantial proportion of people in America.” The firm processes affected person insurance coverage and billing for hundreds of hospitals, pharmacies and medical practices and has entry to large quantities of well being info on a few third of all Americans.
News
The Department of Justice vs. Adobe: The U.S. Department of Justice filed a lawsuit towards Adobe alleging that the corporate hides termination charges and makes it troublesome to cancel subscriptions. Read extra
OpenAI acquires Rockset: OpenAI introduced that it has acquired Rockset, which builds instruments to drive real-time search and information analytics, as the corporate continues to spend money on its enterprise gross sales and tech orgs. Read extra
Buttons are again: Clicks launched a nostalgic, BlackBerry-esque cellphone case that provides a keyboard with bodily buttons to the underside of your iPhone. We received to attempt one for ourselves. Read extra
Where people and AI coexist: Butterflies is a social community the place people and AI work together with one another by posts, feedback and DMs in an effort to have extra inventive relationships with AI. Read extra
Apple kills Pay Later: After launching in late March 2023, Apple’s Pay Later function is not any extra. Instead, Apple Pay customers will be capable of entry loans by a partnership with the third-party app Affirm. Read extra
Beware, Outlook customers: A researcher has discovered a bug that permits anybody to impersonate Microsoft company electronic mail accounts, making phishing makes an attempt look credible and extra prone to trick their targets. Read extra
Perplexity takes on Google: The AI-powered search startup now shows outcomes for factual queries similar to climate and time at a spot, forex conversion, and solutions to basic math queries immediately by playing cards. Read extra
Runway unveils Gen-3: The firm’s newest AI mannequin for producing movies delivers a “major” enchancment in velocity — in addition to extra management over the construction, type and movement of the generated movies. Read extra
Analysis
What ought to AI appear to be?: From black holes to colourful blobs, representing AI in consumer interfaces generally is a problem. While approaches differ to branding purportedly all-seeing, all-knowing, all-doing intelligence, Devin Coldewey explores how corporations have coalesced round the concept that the avatar of AI ought to be non-threatening, summary, however comparatively easy and non-anthropomorphic. Read extra
Why Fisker failed: As Fisker information for Chapter 11 chapter safety, many are left questioning what’s subsequent for the ill-fated EV startup. Sean O’Kane argues that no matter occurs to Fisker or its property, it received’t change the elemental downside: that it wasn’t able to grapple with bringing a flawed automotive to market. Read extra
Pushing ChatGPT’s cultural limits: The present ChatGPT affords solutions which can be too generalized for particular questions that cater to sure communities, as its coaching seems Eurocentric and Western in its bias. With most AI fashions not constructed with individuals of coloration in thoughts, Dominic-Madori Davis and Tage Kene-Okafor report on the Black-owned chatbots and ChatGPT variations that cater particularly to Black and brown communities — and assist founders capitalize on OpenAI’s cultural slip. Read extra