In a kinda-but-not-really surprising announcement, Microsoft unveiled its new division: Microsoft AI. 

In an acquihire of sorts, Microsoft hired the top two in command at Inflection AI, makers of the popular Pi chatbot and Inflection LLM: Mustafa Suleyman and Karén Simonyan.

In a blog post announcing the move, the company named Mustafa as the CEO of Microsoft AI and Karén as the chief scientist. 

What it means: The move was also a bit concerning, given reported concerns about Mustafa’s management style. Also, with Mustafa being the former Co-founder of Google Deepmind as well, you’ve gotta think that Micorosoft has something more up its sleeve in regards to AI. (More on that below.) 

There’s been a lot of talk that Microsoft is TOO reliant on AI startup OpenAI. 

And given Microsoft’s popular Copilot is largely based off GPT-4, it’s no wonder that Microsoft has been scrutinized for not investing more in its own models. 

Reportedly, Microsoft is a 49% owner in OpenAI after more than $10 billion in investments over the years. 

So, as #1 above unfolded, some previous comments from Microsoft CEO Satya Nadella resurfaced with some fresh context: 

In a previous interview with New York Magazine, Nadella said, “"It wouldn't matter if OpenAI disappeared tomorrow. We have the data, IP rights, and all the capability."

What it means: These comments aren’t new. However, with the recent announcement of Microsoft AI and the acquiring of Mustafa Suleyman and Karén Simonyan (see #1 above), these older comments are breathing new life. 

Might we expect to see Microsoft working on its own models this year to reduce its reliance on OpenAI? 

We wouldn’t be surprised.

Welp. That’s pretty cool.

Elon Musk's brain-chip company Neuralink has successfully demonstrated a patient moving a cursor using their brain implant, with the goal of connecting human brains to computers to help tackle neurological conditions.

According to a video released by the Neuralink team, The first patient to receive Neuralink's chip implant, Noland Arbaugh, was able to play chess and a video game using the device.

What it means: This sure seems like the future of brain implants and medical purposes, as science fictiony as that intersection may sound. And if these Neuralink advancements continue, the future could be very bright for people suffering from paralysis or other neurological conditions. 

The bigger picture here? Elon Musk has previously talked about combining the capabilities of Neuralink with Tesla’s humanoid robot, Optimus. We hope it’s mainly for positive use-cases, like Noland’s.

More shakeups in the AI startup space? Yup. 

Stability AI's founder and CEO Emad Mostaque has stepped down from his role and the company's board, with the COO and CTO stepping in as interim co-CEOs.

Emad said he’s going to pursue ‘decentralized AI’ in his future endeavors. 

The startup has faced challenges, including high turnover and unsuccessful fundraising attempts, despite its popular image generation tool Stable Diffusion.

What it means: Could Stability AI be the first big AI domino to fall? Maybe. We talked about this on today’s show, but today’s Unicorns are finding out quickly in 2024 that they have to either fundraise billions or quickly capitalize off their compute. Even though Stability’s product is amongst industry leaders, it seems the money wasn’t reportedly falling into place. 

There’s another entry in the GenAI-powered hardware market: Open Interpreter. Open Interpreter teased their newest hardware gadget powered by AI, the 01.

How’s it work? 

Well, you talk to a $99 ‘button’ and it can interact (via AI, of course) with the web and/or with your computer. All remotely.  

Here’s the demo: 

Why it matters: We previously talked about the Rabbit R1, and Humane’s ai pin. Personally, the Open Interpreter 01 seems to potentially have broader appeal, especially given the fact you can train the 01 on any task on your computer and then have it perform said tasks for you. 

The future of AI software/hardware? Could be.

No, GPT-5 has not been officially announced yet, nor does it have a release date, despite the rumors that it’s swirling out in the wild. 

The OpenAI CEO did, though, talk about the future of the GPT technology on Lex Fridman’s podcast.

What it means: 

Some of our takeaways from this conversation? 

• Altman says GPT-4 “kind of sucks.”

• Altman did say OpenAI would release a new model this year, but didn’t specify it would be GPT-5. (Could 4.5 be first, or will it go straight to 5?) 

• Altman did say that in the “coming months” OpenAI would release “many different things.”

Even before the average person can get their hands on OpenAI’s text-to-video model Sora, apparently the company is shopping it around to Hollywood production studios. 

OpenAI is actively engaging with Hollywood studios, media executives, and talent agencies to promote its AI video generator, Sora according to reports. And apparently, even its CEO Sam Altman is personally meeting with film studios and directors. 

What it means: Two issues here. One – Sora is about SO much more than text-to-video generation. We covered that here. Second: AI videos are 100% going mainstream. Whether you want to believe it or not, the future of cinema has AI playing the leading role. And Sora is likely the first platform that will be debuting everywhere.