AI News | SF Hackathon for AI Agents, Jimmy Apples, AI and mental models, Dr Jim Fan, Max Tegmark.

Today:

AI News | SF Hackathon for AI Agents, Jimmy Apples, AI and mental models, Dr Jim Fan, Max Tegmark.

The SF Hackathon brought together AI enthusiasts from around the world to create and test cutting-edge AI agents. Finalists included Auto Arena, where you can pit autonomous AI agents against web navigation tests, and Stim, which listens to real-time events. There's also Farmers AI, mimicking redditors, and Spin-up agents for community engagement, even after getting banned by Reddit.

AI's power goes beyond simple models, similar to the human brain. Dr. Fan discussed various AI models and their potential, such as Minecraft Voyager, which uses World Modeling for decision-making. Jimmy Apples, known for accurate predictions, claimed OpenAI achieved AGI internally and made predictions about future developments.

WATCH THE VIDEO ON YOUTUBE

ChatGPT use shows that the grant-application system is broken

Navigating the slog of grant applications is a notorious pain for scientists: a cocktail of endless docs, tight deadlines, and sky-high rejection rates. While the laborious process aims to sift out the non-serious applicants, it eats up a chunk of valuable research time.

Now, with AI like ChatGPT shouldering some of the burden, questions are bubbling up: if a bot can ace writing those tedious grant docs, why make scientists do it? The process is proving to be an exhausting ritual rather than a valuable filter for quality, poking holes in the traditional grant-application system’s rationale.

READ THE ARTICLE ON NATURE

OpenAI Just Replaced It’s “Core Values” With Completely Different Ones

OpenAI, known for its highfalutin work in artificial intelligence, recently did a switcheroo on its core values, making "AGI Focus" a top priority. AGI, or Artificial General Intelligence, means building systems either "smarter than humans" or on par with average human intelligence—depends on which OpenAI chat you're tuning into.

While once flaunting values like being "Audacious" and "Impact-driven," they’ve shifted gears to spotlight values like being "Intense and scrappy" and working to "Make something people love." The move raises eyebrows: Can values be "core" if they're swapped out like seasonal wardrobes?

Plus, OpenAI, originally birthed as a nonprofit by bigwigs like Elon Musk, seems to drift from its initial “AI-for-good” mojo, transitioning to a profit-making machine, causing some, like Musk, to jump ship. While they’re still tipping their hat to building beneficial and safe AI, one can’t help but side-eye the fuzzy, flip-floppy signals coming from the firm’s corner.

READ THE ARTICLE ON FUTURISM

AI Chip Reseller Raises $50 Million in Financing Led By Kleiner Perkins

Together, a fresh startup that rents out fancy computer chips (made by Nvidia) for AI work, just bagged a cool $50 million in funding led by Kleiner Perkins. Why? Because artificial intelligence, like the tech running ChatGPT, is hungry for these special chips.

The San Francisco gang at Together don’t just rent out the chips through the cloud; they also sell software that makes it a breeze for other companies to use these brainy language models. This deal not only shows the sizzling demand for AI tech but also pops Together’s value above $300 million. Even Nvidia, the chip-maker itself, might throw some dollars into this funding hat.

READ THE ARTICLE ON THE INFORMATION

AI forces a rethink on executive MBA teaching

Artificial Intelligence (AI) is shaking things up in business schools, especially in Executive MBA (EMBA) programs. The AI technology, like ChatGPT, has proven its chops by passing MBA entrance tests and even snagging a B grade in an MBA course.

This wake-up call has sent business school bigwigs scurrying to mesh AI into their EMBA curricula. Schools like Esade and Imperial College Business School are diving deep, offering students real-world AI experiences and even utilizing AI to shake up their teaching and assessment strategies.

This ain't just about learning the tech, but figuring out how it slots into business strategies and leadership. Richard Manga, for example, hit the jackpot in landing his dream job after juicing up his AI know-how during his EMBA at HEC Paris.

READ THE ARTICLE ON FINANCIAL TIMES

AI just got 100-fold more energy efficient

Scientists at Northwestern University hit a home run in the AI field, creating a tiny, super energy-efficient device that can sort out data and perform AI tasks without needing to bounce information to the cloud. Imagine your smartwatch getting 100 times better at using its battery while also being able to tell you real-time health info, like spotting a funky heartbeat on the spot with nearly 95% accuracy.

The wizards crafted this gadget using some fancy materials like molybdenum disulfide and carbon nanotubes instead of the usual silicon. This lets it quickly swap through different data processing steps using just two devices instead of over 100 transistors, saving a ton of energy and space. Besides giving you instant health insights without sending your data on a round trip to the cloud, it's a solid win for keeping your info safe and sound from prying eyes.

READ THE ARTICLE ON NORTHWESTERN

Millions of Workers Are Training AI Models for Pennies

In 2016, Oskarina Fuentes from Venezuela started tagging data for AI algorithms with Appen, an Australian company, amid her country's economic collapse. Think about those "identify the traffic light" tasks we all do online; well, companies like Amazon and Google use loads of those, prepped by folks in cheap labor countries. The money involved? The global gig for collecting and tagging data was worth $2.22 billion in 2022 and could hit $17.1 billion by 2030.

Turns out, it's a trend. Workers in places like India, Kenya, and even refugee camps are doing these micro-jobs. Some experts think this is just digital colonialism. Workers from the Global South prep data for AI that the richer North uses. Plus, these jobs might not stick around. Today's hot skill in AI tagging might be old news tomorrow.

READ THE ARTICLE ON WIRED

How generative AI is creeping into EV battery development

If you're into batteries, finding the right combo of these molecules can give us some wicked good batteries for cars, the power grid, or even electric planes. But here's the kicker, it's like finding a needle in a haystack. It could take ages and a ton of tries. Enter Aionics, a startup that reckons their fancy AI can speed up this wild goose chase. Their game? Focusing on the middle part of the battery, the electrolyte.

Using some heavy-duty AI, they're running tests on a crazy number of molecules in a flash, trying to find that sweet spot. They're also trying something called "generative AI". Basically, instead of just testing what we already know, they’re dreaming up entirely new molecules for specific tasks.

Plus, they're taking some tools from Carnegie Mellon University and using a chatbot from OpenAI (that's some next-level AI stuff) to help their scientists sort things out faster.

READ THE ARTICLE ON TECHCRUNCH

Burned-out parents seek help from a new ally: ChatGPT

Parents are now leaning on ChatGPT as a handy helper for raising their kids. It's become a tool for the frazzled folks trying to juggle it all. From quickly planning mermaid-themed bashes to answering the age-old "Why's the sky blue?", ChatGPT's got your back.

Some even use it for bigger moments, like scripting heart-to-hearts with their teens. Keith Foxx, a dad, uses it to send his high schooler uplifting messages. Foxx's counter? He's just trying to connect, and his kid's in on the joke. In the end, it's cool tech, but not a magic fix. Use wisely.

READ THE ARTICLE ON AXIOS