Elon grabs $6 billion for AI, promises “Gigafactory of Compute”, makes bank

Like him or not (we don’t), you have to admit that Elon Musk has a knack for staying in the public eye. This week, the world’s richest man-baby pulled off a trifecta of headline news events. First, he managed to convince some of the smartest and richest investors in Silicon Valley (plus the Saudis) to pour $6 billion (with a “b”) into his AI startup, xAI, for only 25% of the equity. Second, he promised to use the money to build the world’s largest AI data center, nicknamed the “Gigafactory of Compute” with partner Oracle Systems. Third, he got his pals on the Board of Tesla to vote him the same $56 billion pay package that was recently struck down as excessive by a court ruling on a shareholder lawsuit. A busy week, even for Elon.

Elon raises $6 billion for AI…

… and promises “Gigafactory of Compute”…

… while getting the Tesla Board to vote him a $56 billion pay package

Nvidia CEO may be on track to become richest man in the world

If Elon is tech’s bad-boy man-baby, Nvidia CEO Jensen Huang, gives off hard-working immigrant Asian Dad energy. Huang immigrated to this country with next to nothing, cleaned toilets for a living early on, then started a gaming computer company that almost went belly up with its first, failed product. Fighting back from this early failure, Huang has built the third most valuable company on the planet, after Microsoft and Apple. His personal stake is estimated at $91 billion. This is almost half of Elon Musk’s $210 billion and Jeff Bezos’ $200 billion. With Nvidia still commanding 95% of the AI chip market, Nvidia stock may have a way to go still.

Nvidia CEO may be on track to become the world’s richest man

OpenAI announces new Safety Team

A week after OpenAI dismantled its former Safety Team (which was charged with making sure that any AI released by the company wouldn’t go rogue and exterminate humanity), OpenAI announced its ALL NEW Safety Team, which was… (wait for it) CEO Sam Altman and 2 other Board members. (In other news, the Fox was put in charge of the Henhouse.) So drama continues at OpenAI, as its idealistic Old Guard battles against the new management which wants to commercialize the company’s AI models as quickly and profitably as possible. One former Board member, who was ousted after trying to fire Altman last November, went public this week with more potshots at Altman’s leadership.

OpenAI announces new safety team

Former OpenAI Board member spills the tea on November’s coup

Smartphone app “Nanni AI” decodes baby cries with AI

Startup Ubenwa Health has developed a smartphone app that can listen to a baby’s cries, identify the likely cause (hunger, pain, etc.), and give suggestions for how best to soothe the crying infant. The company’s founders trained an AI model on thousands of audio clips of babies crying, with annotation as to the probable cause. The app has been downloaded some 150,00 times, and the AI model is being continuously updated as parents use the app with their own babies. The goal is to improve the sleep and wellbeing of both the babies and the parents.

Smartphone app Nanni AI decodes baby cries with AI

YouTube Music app lets you hum that song you’re searching for

The app Shazam has been helping users identify half-remembered songs with hums and whistles for over 20 years, using old-school algorithms. The YouTube Music smartphone app has given this “Hum to Search” function an AI update, and the early reviews are good.

YouTube Music app lets you hum that song you’re searching for

Canva design app gets AI remake

Canva, the popular user-friendly design app for users who don’t need or want the complexity of Photoshop, has gotten an AI makeover. Image creation, image editing, copy writing, and more have been given a ChatGPT-type interface where you can just tell Canva what you want to create, and it will execute your instructions. You can even create full-featured courses of instructional material. Meanwhile, Adobe is also putting AI into its flagship Photoshop, and developing an AI-first image creation product known as Adobe Firefly.

Canva design app gets AI remake.

Researchers create multilingual text-to-sign language app

AI is very good at translating text from one language to another. Now researchers have developed an AI model that translates text from a variety of languages to one of several international sign languages for the deaf and hard of hearing. Once the text is entered, a video clip of an AI avatar pops up and performs the sequence of signs that express the meaning of the text. This opens up greater opportunities for communicating between hearing and non-hearing individuals.

Researchers create multilingual text-to-sign language

Colorado school district uses AI to solve bus driver shortage

Since the COVID pandemic, there has been a growing shortage of school bus drivers nationwide. One school district in Colorado faced such a critical shortage of willing bus drivers, able to hire only half as many as the projected need, that they threw traditional solutions out the window and turned to an outside rideshare company. Using the rideshare company’s sophisticated AI routing software, the school district was able to find optimal routes for the buses to serve the maximal number of children with the fewest number of drivers. Any holes in the routes were patched over with rideshare trips with specially-trained drivers who underwent rigorous background trips. The total cost of the solution was hundreds of thousands of dollars less than the traditional way of doing things, and parents were texted when their child was picked up, and again when the child was dropped off, a much-appreciated extra level of service.

Colorado school district uses AI to solve bus driver shortage

AI pierces the secret life of plants

Global agriculture is facing a crisis - a growing world population to feed, while climate change threatens crop yields. AI is poised to help in major ways. Recent advances in AI gene decoding, gene manipulation, and protein design promise to speed up development of hardier and more productive crop varieties. AI systems are also being developed to capture real time data on soil conditions, rainfall, pests, and a myriad of other factors affecting crop yield, and produce continuously updated optimal cultivation plans for each particular plot of land throughout the growing season. And finally, GPS guided precision robo-tractors can plant and harvest crops at a density and uniformity unattainable by unaided humans.

AI pierces the secret life of plants

AI augments cardiac caths with real time performance data

Researchers at the Mayo Clinic have developed an AI system that addresses one of the longstanding conundrums of cardiac angiography - in the cath lab, the physician has excellent visualization of the cardiac anatomy, but only an impressionistic idea of vital parameters such as cardiac index or filling pressures. The Mayo Clinic team trained an AI model on 20,000 angiography videos, paired with echocardiograms or right heart catheterization of the same patients. The model can predict the performance parameters from the angiography video alone, with an accuracy of around 80%. Mayo is continuing to refine the model before deployment into clinical practice.

AI augments cardiac caths with real time performance data

AI decodes fruit fly vision, giving clues about human sight

A young Cold Spring Harbor Laboratory scientist developed an AI model of the fruit fly brain - and made a striking discovery. The mapping of behavior from visual stimuli involved complex webs of neurons, looking more like dense subway maps rather than the mostly linear input-output architecture expected. Researchers at CSHL theorize that neuronal webs for vision in the human brain, with 1,000,000 times as many neurons as the fruit fly, may be correspondingly more complex.

AI decodes fruit fly vision, giving clues about human sight