Elon Musk’s newly announced AI company, XAI, its groundbreaking mission to understand the universe, its illustrious team, and how this might transform our understanding of artificial intelligence and AI safety.
Elon Musk announced XAI – the answer to OpenAI = X.AI
Elon Musk, well-known for his ventures Tesla and SpaceX, has introduced X.AI, a new artificial intelligence (AI) player aiming to understand reality. His latest endeavor aims to compete directly with OpenAI’s Chat GPT. Musk’s X.AI is staffed by big shots from Google DeepMind, Microsoft Research, and OpenAI itself.
Launched on July 12, 2023, the ultimate mission of X.AI is rather unique: to comprehend the universe’s true nature. The idea is that an AI focusing on understanding the cosmos is less likely to obliterate humans because we’re part of the fascinating universe it’s studying.
The star-studded team, led by Musk, boasts of experienced researchers from leading tech giants, contributing to breakthroughs like Alpha star, Alpha code, Inception, Minerva, GPT 3.5, and GPT4. The advisory board includes the Director of Center for AI Safety, Dan Hendricks, emphasizing Musk’s worry about AI’s potential existential threat.
However, this move is not without controversy. Musk’s harsh criticisms of OpenAI have led to public disagreements and a growing rift in the AI community. Nevertheless, X.AI, backed by Musk’s wealth and influence, is shaping up to be a major player in the AI world.
Elon Musk says Tesla will spend $1 billion to build a ‘Dojo’ A.I. supercomputer
Elon Musk’s Tesla is shelling out over $1 billion to create a mega-powerful AI computer named Dojo due to a shortage of Nvidia’s high-tech chips. Tesla is super hungry for these chips to improve their Full Self-Driving (FSD) software, which currently goes for a cool $15k, and to boost their Optimus humanoid robot project.
Turns out, Nvidia can’t supply Tesla with enough chips ’cause they’ve got customers coming out of the woodwork. Musk isn’t even sure if Dojo will beat out just buying Nvidia chips, but he’s gung-ho on trying.
By next year, Tesla’s goal is to achieve 100 exaFLOPS (that’s a mind-boggling level of computer performance) with Dojo, which would catapult Tesla into the top five global providers.
Tesla’s approach to getting this done is unique. Rather than spending a fortune on numerous sensor systems like other companies do, Tesla aims to make their cars smart enough to drive anywhere, all on their own, just by using camera data. To train them, Tesla is using real-world driving data from its customers, which will be fed to Dojo. This new AI supercomputer will lower the cost of AI training using custom-built tech developed in-house at Tesla. It’s designed for processing video data instead of using language like other AI models.
GitHub’s Copilot Chat AI feature is now available in public beta
GitHub has tossed its new tool, Copilot Chat, into the public beta ring. This shiny new feature is like a personal assistant for developers, aiming to help them get through coding faster and easier. Think of it as a turbo boost for coding, available to all business users through Microsoft’s Visual Studio and Visual Studio Code apps.
Copilot Chat, born in March as the main deal of GitHub’s Copilot X plan, grew from the original Copilot code tool, but got some muscles from OpenAI’s GPT-4 model. What’s the big idea here? This AI helper is built to make developers’ lives a breeze by getting some of the trickiest tasks done with just a few prompts.
This tool has a good sense of what’s going on in the code, and error messages, and offers support that’s tailored to the project at hand. From real-time guidance, explaining suggestions, and troubleshooting, Copilot Chat is there to lend a hand.
OpenAI Backs Idea of Requiring Licenses for Advanced AI Systems
OpenAI, a big-time AI company, put out a call for government licenses for advanced AI tech. This came after their top dog, Sam Altman, had a chat with folks at the White House about AI rules.
Altman’s idea? Create a new government group to give the green light to high-level AI tech. This group could also pull the plug if the AI doesn’t meet safety rules. Plus, Altman underscored the need for outside checks to make sure AI plays by the rules before it’s let loose on the public.
Yet, Altman realizes not all AI should face the same tough standards. He thinks Congress should set specific goals for what kind of AI needs a license. Like, for instance, AI that can change people’s minds or create new living organisms.
TSMC delays Arizona factory that will eventually build chips for iPhones and AI
TSMC, the world’s biggest chipmaker, is hitting the brakes on its new factory in Phoenix, Arizona, pushing the start of production from 2024 to 2025. Why? They simply can’t find enough skilled workers in the US to get things rolling. Tech titans like Apple, Nvidia, and AMD are all waiting in line to get chips from this factory, but they’ll have to hold their horses.
Originally, TSMC planned to start churning out 4nm chips next year at this factory. But with the worker shortage, they’re now sending over a “task force” of 500 technicians from Taiwan to help set things up and train the local crew. A second factory, planned for 2026, will produce even more intricate 3nm chips.
Meta launches an AI research community but devotes few resources to it
Meta is stepping up its AI game by starting a new group, the Open Innovation AI Research Community. The goal? Spark more openness, innovation, and teamwork in AI research, particularly when it comes to making AI models like ChatGPT safer and more private.
Meta says the group will be run by its members, with the company’s AI crew playing a backseat role as a “facilitator.” The idea is to get a bunch of smarty-pants from the AI world working together on big, open-source projects and training the next-gen of AI pros.
And who’s footing the bill for all this research? Looks like it’s up to the group members themselves, as Meta didn’t mention allocating any cash or computer resources for the project. That’s a tall order considering AI research ain’t cheap.
They’ve got until September 10 to sign up for the Open Innovation AI Research Community. Meta says they’re open to all kinds of research folk and that more than one person from the same university can apply. But only time will tell if this move is a game-changer or just a PR stunt.
Microsoft’s Copilot has Wall Street more bullish on AI prospects
Microsoft’s pricing for its new AI-integrated product, M365 Copilot, has exceeded Wall Street analysts’ expectations, boosting the company’s prospects. At $30 per month, UBS analyst Karl Keirstead highlighted that this was at the upper limit of projected pricing.
The M365 Copilot is an AI-powered chatbot able to perform a variety of functions, such as summarizing unread emails, reformating PowerPoint presentations, and drafting documents on demand. This product could potentially add an estimated $7-9 billion to Microsoft’s Office revenue by FY25, according to Keirstead, which is a considerable increase from the previously estimated revenue of $3-5 billion.
AI has been a significant factor in Microsoft’s nearly 50% stock increase this year. Earlier this year, Microsoft announced a $10 billion investment in OpenAI, the parent company of the popular AI chatbot, ChatGPT. Microsoft also hinted at a “new AI-Powered Microsoft Bing” in February, further intensifying the competition in the AI space.
This trend in AI development and integration has led analysts to increase their projections and investors to shift their focus from concerns about slowing cloud growth. The Bank of America Research analyst Brad Sills raised his price target on Microsoft, viewing Microsoft as a leading AI play in software.
SAP invests in generative AI startups Anthropic, Cohere and Aleph Alpha
SAP, the bigwig consulting firm from Germany, is putting its money into generative AI, splashing out on three startups: Anthropic, Cohere, and Aleph Alpha. The deal’s specifics? Kept under wraps, but it’s all part of SAP’s over $1 billion bet on AI-backed business tech via Sapphire Ventures.
SAP’s strategy chief, Sebastian Steinhaeuser, reckons we’re at a turning point with generative AI on the brink of shaking up how businesses run. SAP is dead set on forging an AI ecosystem that jives with their top-notch business apps, helping their clients go full throttle.
All three investments line up snug with SAP’s interests. Anthropic is working on Claude, an AI system built for business tasks like generating answers, coding, and workflow automation. Cohere provides a text-generating platform via an API, catering to companies’ data needs. Aleph Alpha, already in SAP’s good books, develops and hosts multi-modal, multi-language models with an emphasis on data privacy and security.
More than 1,300 experts call AI a force for good
Over 1,300 professionals are stepping up and claiming artificial intelligence (AI) ain’t the bad guy. The UK’s IT powerhouse, the BCS, orchestrated this. They’re hoping to squash the fear of a robot uprising. Even though big names like Elon Musk have paused their AI development due to concerns about super smart AI causing trouble, the BCS crew doesn’t think we’re close to that.
One of the letter’s supporters, Richard Carter, founder of an AI cybersecurity startup, thinks the scary predictions are out of touch. They don’t think AI can pose a risk to humanity just yet.
Folks who put their names on the BCS letter come from all walks of life. While they may not be as famous as Musk, they’re still serious about AI’s potential benefits. For instance, Hema Purohit, an expert in digital health, said AI could spot severe health issues faster and streamline testing for new medications.
The letter also suggests that the UK could play a major role in defining AI standards and guidelines. By doing so, it could become a hotspot for ethical, inclusive AI. UK Prime Minister Rishi Sunak plans to host a global summit on AI regulations soon.
Neura Robotics picks up $55M to ramp up in cognitive robotics
Neura Robotics, a German outfit that crafts thinking robots, has raised a hefty $55 million. Their robots can remember, tackle complex situations, and work with humans. The company plans to use these fresh funds for research and expansion in Asia and the U.S., as well as boosting production. They already have orders worth $450 million lined up for the next five years.
The company has developed three types of robots – MAV mobile robot, LARA high-end “cobot”, and MAiRA, the first “cognitive robot”. Their clients use these robots for various industrial applications. Their upcoming service robot MiPA, which will assist in offices, care facilities, and homes, is priced to be affordable.
Neura’s robots are equipped with the ability to be trained and operated in any language and dialect, and they can work both online and offline. They are designed to work with humans and have safety features to adjust their movements in response to human contact.
Neura’s grand plan is to control the entire package – both the software and the hardware – and work closely with customers to develop specialized apps in areas like welding, warehousing, gluing, sanding, and assembly. In the future, Neura hopes to step directly into the consumer market with its robots.
Ukrainian-founded Duolingo rival Preply banks $70 million to push into A.I.
Preply, a US-based language learning service with Ukrainian roots, recently nabbed $70 million to beef up its AI capabilities. The company, which teams folks up with human tutors to learn new languages, has raked in a total of over $170 million in funds to date.
Preply’s not just a hit with your everyday language learner, big businesses like Datadog, GroupM and Bain also shell out for its services to buff up their teams’ foreign language skills. Though still not turning a profit, Preply’s seen its revenues multiply by ten in the past three years.
The new funding will help Preply take a deeper dive into AI, adding features like an AI assistant to help tutors plan lessons. The move comes as Preply’s competitors like Duolingo also invest in AI to personalize learning experiences.