Welcome, humans.

Ahhhh, the morning post-AI Christmas is finally here. The last two days were more exhausting than being at a Christmas holiday stone-cold sober. We might be reeling from a food AI coma, but hey, now we get to mess around with our new gadgets and decide which gifts are keepers.

Here’s what you need to know about AI today:

• Free and paid users receive 16 and 80 GPT-4o messages every 3 hours, respectively.

• Yet another leading AI researcher has stepped down from OpenAI.

• Microsoft’s carbon emissions have increased 30% since 2020 in part due to AI.

• Try to spot which image was generated by Imagen 3!

On the podcast: Pete’s analysis of Google’s major AI launches and how Google stacks up against the competition. (Apple Podcasts, Spotify, YouTube).

A 60-second rundown of the last two days.

A TON of AI stuff went down over the past few days, but TBH, many of the headlines aren’t relevant for everyday users—they’re futuristic demos or niche developer tools.

Here’s what actually matters from the past two days:

1. The biggest unlock: Access to ChatGPT-4o has been or will soon be granted to everyone. Our tests show it’s a noticeable improvement over GPT-4 and Claude 3 Opus.

• Free users get 16 ChatGPT-4o messages per 3 hours.

• Plus users get 80 ChatGPT-4o messages per 3 hours

• Teams users 160 ChatGPT-4o messages per 3 hours.

So only pay for ChatGPT+ if you’re sending more than 16 messages/day to ChatGPT-4o. We’ve heard rumors that OpenAI might announce another major update soon (perhaps a search feature?), but for now, ChatGPT+ users simply get increased usage.

2. ChatGPT Voice Mode, the talking ChatGPT that everyone’s excited/nervous about, will initially roll out to Plus users first over the coming weeks.

3. Google’s latest model, Gemini 1.5 Pro, which lets you upload 1,500 pages of text, is coming to Gemini Advanced (which costs $20/month).

This tool is perfect if you need to summarize or talk to massive documents—it’s 99% accurate at uncovering obscure facts buried in heaps of text (the LLMTest_NeedleInAHaystack test).

Then there’s using Gemini 1.5 Pro in Gmail, Docs, Sheets, and Slides. In theory, these tools have big potential, but it's TBD if they perform as well as their demos suggest.

You’ll need an existing Workspace plan for this one.

4. Finally, Google AI search is here. If you’re in the US, you’ll start seeing more AI summaries in your searches. It’s coming to 1B users by EOY.

We think Google SGE’s AI answers are excellent for quick info, but some reports suggest they hallucinate a decent amount, so proceed with caution.

FROM OUR PARTNERS

Close more deals & make more money with this magical AI tool. 

At The Neuron, we've already automated big chunks of our sales process using AI.

If you want to do the same and make more money for your business, try Attention ASAP. 

It’s an AI sales genius that:

1. Listens to your sales calls and reveals what your customers are really thinking.

2. Automatically updates your CRM based on sales calls.

3. Sends follow-up emails in under a minute.

It's magic—no joke. Check out their 45-second demo here. 

Join 100+ sales teams and start sealing more deals with Attention. 

Around the Horn.

• A top OpenAI superintelligence researcher has stepped down following Ilya Sutskever’s exit.

• Anthropic hired one of Instagram’s cofounders as its new Chief Product Officer.

• Google is building an Android feature that identifies scam calls and notifies you instantly—it’s powered by Gemini Nano.

• Microsoft’s carbon emissions have increased by 30% since 2020 in part due to additional data centers for AI.

• Stability AI is talking to potential buyers.

Treats To Try.

1. *Speechmatics, a leading AI speech recognition company, just launched Audio Events, a new feature that automatically detects and labels nonverbal sounds like [music], [laughter], and [applause] in audio. It’s perfect for media, education, customer service and more. Get early access to Audio Events here!

2. Chatter is a new interactive AI podcast (ask "what's happening in politics today") powered by Hume's voice API that can scour the web.

3. Anthropic Console has a new tool that generates prompts optimized for various tasks.

4. Voicenotes is an AI notetaker that organizes your thoughts.

5. Glitter AI converts your mouse clicks + voice into a written step-by-step guide with screenshots that you can share with colleagues.

*This is sponsored content. Advertise in The Neuron here.

Thursday Trivia.

Google’s newest image model, Imagen 3, is likely the second tool after Midjourney capable of creating faces that look and feel 100% real. Scary stuff.

One glance, everyone knows the rules: one image is AI-generated, and the other is real.

Which is which?

A.

B.

A Cat's Commentary.

Trivia Answer: A is Imagen 3 and B is a real photo.