Jordan almost had to talk at 2X speed today to get to all the AI news. 

Sheeeeesh. 

We've got government vetoes, purported price hikes, AIs fact checking AIs and enough OpenAI drama to fill a Silicon Valley soap opera. 

Multiple seasons. 

We’ve got the entire lowdown for ya right here, as we do every Monday. 

Each week, we bring you the AI news that matters. So don’t get lost trying to connect dots between all the comings and goings of AI, and lettuce be your guide. 

Alright. Let’s dive in. This week’s a doozy and a half y’all. 

Governor Gavin Newsom axed California’s SB 1047, a bill that would've required AI companies to assess their systems for bias and discrimination before release. 

The bill would have held tech companies legally accountable for harms caused by AI models and required a "kill switch" for AI systems in case of misuse. Major tech players, including OpenAI and Andreessen Horowitz, lobbied against the bill.

The bill aimed to create accountability in AI development, but Newsom argued it could actually stifle innovation and create a patchwork of state regulations. This decision comes as the U.S. grapples with balancing AI advancement and potential risks.

What it means: 

This is just Round 1. 

Let’s be honest — Newsom had two very strong factions tugging him in opposite directions. Hollywood (and legislators who passed the bill) strict regulations while Big Tech thought it was too restrictive. 

Newsom's veto signals that unrestricted AI development trumps immediate safety concerns. Newsom, though, still highlighted the good in the AI bill while vetoing it, a clear signal that we’ll likely be seeing a revised version of this bill soon. Newsom said that 32 of the world's 50 leading AI companies call Cali home, so any harsh state legislation could actually hinder the world’s innovation. 

And anything that actually passes in California could have long-term implications on the U.S. economy. 

Bills, bills, bills. 

According to reports, OpenAI is set to increase ChatGPT Plus prices from $20 to potentially up to $44 a month by 2029. 

This comes as reports surface of the company burning through $700,000 daily – that's a staggering $255.5 million annually. Despite boasting 100 million weekly active users, OpenAI is still operating at a loss, raising questions about the sustainability of current AI business models.

What it means: 

The AI gold rush is hitting its first major reality check. This price hike exposes a harsh truth: cutting-edge AI is insanely expensive to run.

Expect a domino effect across the industry. Small businesses relying on OpenAI's tech might get priced out, potentially stifling the very innovation everyone's betting on.

Buuuuuuuuut, let’s be real y’all. 

We’ve been spoiled with $20 a month access to these LLMs. In time-savings alone, even mediocre models are worth their weight in gold if you actually know what you’re doing. 

Apple has reportedly tabled its potential $1 billion investment in OpenAI after months of talks. 

This surprising move comes as OpenAI is set to soon close a record investment round of up to $6.5 billion.

Apple's decision to back out of such a high-profile investment at the last hour has sparked speculation about its own AI strategy and views on OpenAI's future.

Other tech giants like Microsoft and Nvidia remain in discussions to invest, with Microsoft expected to contribute another $1 billion after already investing $13 billion in OpenAI.

What it means: 

The initial reporting that Apple was even involved in the funding discussions was pretty eye-opening. 

Given that Apple rarely invests in third-party tech companies, it was pretty telling that Apple was set to write some big checks.

Pulling out at the last minute, though, is definitely sending mixed messages. 

After all, Apple has completely dropped the ball on its slow-rolling Apple Intelligence rollout. They couldn’t handle everything in house, so they formed a VERY public relationship with OpenAI and handed user questions to ChatGPT. 

And, TBH, the initial Apple Intelligence in beta isn’t very impressive. 

Even with this partnership, though, Apple is pulling its investment? 

Huh? 

On one hand, they are publicly relying on OpenAI to handle the AI heavy lifting for its much-hyped Apple Intelligence. Yet on the other hand, they’re pulling investment back? 

Why not pay the piper who’s planting to feed your people? 

Facepalm. 

Seems like Apple is continuing to drop the ball on its AI Go-To-Market strategy.

Yeah. Even more OpenAI news. We said there was drama shorties. 

OpenAI is reportedly undergoing significant leadership changes as three top executives announced their departure on the same day. Chief Technology Officer Mira Murati, Chief Research Officer Bob McGrew, and Vice President of Research Bharat Zoph are all leaving OpenAI.

Murati's exit is particularly notable given her key role in developing technologies like ChatGPT and DALL-E. 

She expressed that her six and a half years at OpenAI were a privilege, stating that the timing felt right for her to step away.

What it means: 

From the surface, it appears OpenAI's facing a slight existential crisis. 

This brain drain could be the canary in the coal mine for deeper issues. 

More on that later in this newsletter. 

(Yes, there’s EVEN MORE OpenAI news. Lolz) 

Llama’s gone multimodal y’all. 

Meta has released Llama 3.2, their first multimodal AI model capable of handling text and images in the input and output. The model comes in four sizes: 1 billion and 3 billion parameters for edge AI, and 11 billion and 90 billion parameters for multimodal AI.

The company also introduced Meta Creator AI and unveiled new AR glasses called Orion.

What it means: 

Meta’s making waves in the Open Source (err… Open Weights?) AI space. They continued their onslaught at their Meta Connect conference, throwing their open weight around at the closed and proprietary model makers like OpenAI, Anthropic and Google. 

As Meta brings its AI to more devices (Ray Ban glasses, phones, Orion) and places (Meta Creator AI), their impact is continually growing. 

If you didn’t already have Meta as a major LLM contender on your 2025 Bingo card, it’s best to pencil them in now. 

And by training on billions of social media users, they’ve got an advantage that the rest of Big Tech can’t (legally) touch.

We gave you a glimpse at the on-device capabilities of Llama 3.2 here. 

The FTC has launched Operation AI Comply, targeting companies that use AI tools to mislead consumers. The agency emphasizes that there is no exemption for AI in existing laws.

Among the companies facing action is Do Not Pay, which has falsely claimed to offer an AI service capable of replacing human lawyers. 

The FTC proposed a settlement requiring it to pay $193,000 and notify customers about the service's limitations.

What it means: 

While we don’t see any new federal laws here in the U.S. regarding AI, that doesn’t mean it’s a complete Wild West. 

Why? 

Cuz certain federal agencies can still hip check companies as they see fit. 

The FTC is taking a strong stance against deceptive AI practices, signaling that companies can't hide behind AI claims to avoid regulatory scrutiny. 

This could reshape how businesses market and implement AI technologies.

Companies are gonna have to think twice before sprinkling some AI on their service and blindly saying their shiny new AI will save you a bajillion dollars. 

AI fact-checking AI. How Meta. 

(Or…. How Microsoft?) 

Microsoft has launched a new AI feature called Correction, aimed at addressing inaccuracies and hallucinations in AI output. 

The feature is available in Azure AI and scans AI outputs for inaccuracies by comparing them against a customer's source material.

While this feature aims to reduce misinformation generated by AI, Microsoft acknowledged that it still may produce errors as it relies on both small and large language models to align outputs with grounding documents.

What it means: 

AI fact-checking AI is gonna continue. Just like how AI models are creating new data for new AI models. 

Inception-y, right?

Microsoft is acknowledging the limitations of current AI systems and taking steps to improve accuracy. 

This move highlights the ongoing challenges in AI reliability and the need for continuous improvement in AI outputs.

Admins, rejoice! 

Using Gemini inside Google Workspace is going to get slightly less confusing soon. Well, reportedly. 

Google has announced that Gemini AI will be integrated directly into Workspace as a standalone feature instead of an add-on. 

This will be included in the Workspace business, enterprise, and frontline plans starting in Q4, eliminating the need for a separate purchase.

Google also updated NotebookLM with the ability to upload audio files, PDFs, and YouTube videos. 

Yassssssss we’re loving this new feature. 

What it means: 

TBH, we both think this is a great move from Google and also a numbers game. 

If your company was using Google Workspace, it was legit a maze and a half to implement Gemini across your organization. So, it’s a smart move to enable Gemini by default, fosho. 

Our take? Google just wants gaudier numbers of business professionals using their tools. 

More people using AI = more data and increased stock price.

So much for that warm feeling of a nonprofit, huh? 

OpenAI is reportedly planning to restructure from a nonprofit to a for-profit corporation. This move would remove the nonprofit board's control over the for-profit entity, potentially making OpenAI more appealing to investors.

CEO Sam Altman could reportedly receive up to 7% equity in the new for-profit structure, which could be valued at $11 billion. Altman called this notion "ludicrous" according to reports.

Can anyone else NOT read the word ludicrous without feeling nostalgic about early 2000s rap or just us? 

What it means: 

Can we finally just call OpenAI a company now? 

Like literally every other tech company. 

OpenAI has had a for-profit subsidiary for years, but it’s prolly best to just cut the nonprofit tape here. 

With a record-breaking funding round about to close, it’s clear that OpenAI has to focus on a profit-first mentality, which is a little weird if you’re still trying to publicly operate as a nonprofit.

Saving the best for last? 

Maybe. 

OpenAI has expanded its advanced voice mode feature to all paid ChatGPT Plus and Teams subscribers, with enterprise and edu customers expected to gain access next week. The new voice mode offers faster latency than a human and sounds like talking to a real person.

Currently, the advanced voice mode doesn't work with other ChatGPT tools like GPTs, file uploads, or Browse with Bing. 

The previously announced video and screen sharing features are still not available.

What it means: 

Does this change how we all work? 

Not now. 

But once AVM gets all the tools (Browse with Bing, GPT access, Advance Data Analysis, etc)…. Watch out! 

We gave you the FULL run-through here. 

Pretty soon, you’ll be able to talk with ChatGPT like you would any other human. You can upload all your important docs, have it browse the web for you, and create docs and visualizations. 

Imagine doing that all on your way to the office in your car. 

It’s like having the world’s smartest assistant doing your daily tasks for you before you even officially clock in.