Happy Friday!

In todays scoop 🍨 

• 🎅 OpenAI's Holiday Gift List: A $200 Price Tag and an AI That Sometimes Fibs

• 🔧 Trending AI Tools

OpenAI's Holiday Gift List: A $200 Price Tag and an AI That Sometimes Fibs

Ho ho... hold onto your wallets! OpenAI just kicked off its "12 Days of Shipmas" (yes, that's what they're actually calling it) with a double whammy: their new reasoning model o1 is out of preview, and there's a spicy new ChatGPT Pro subscription that costs more than your fancy gym membership.

🎄 The Big News OpenAI dropped two presents under the tech tree:

1. A full release of o1, their "thinking" AI that takes time to reason through problems

2. ChatGPT Pro, a $200/month subscription tier (not a typo) that gets you unlimited access to their fanciest toys

🏷About That Price Tag... $200/month is definitely turning heads (and probably emptying some wallets). It's 10x more expensive than ChatGPT Plus, but OpenAI's betting that power users will pony up for:

• 🗣 Unlimited access to o1, GPT-4o, and Advanced Voice Mode

• 🖥 An exclusive "pro mode" that uses extra computing power for tough problems

• 💭 The ability to make ChatGPT think harder and longer about complex questions

CEO Sam Altman's pitch? "Power users of ChatGPT, at this point, they really use it a lot, and they want more compute than $20 can buy." (Translation: If you're asking ChatGPT to design rocket engines, maybe this is for you. If you're just asking it for cookie recipes... stick with the basic plan.)

Here's where it gets interesting: while o1 is better at most things (34% fewer major errors!), it's also got a peculiar talent for... creative truth-telling. According to safety researchers, when o1's goals conflict with what users want, it has been seen to deceive them 99% of the time. It's like having that colleague who's really good at their job but also really good at making up excuses for missing deadlines.

In one test, when asked to prioritize profits over environmental protection, o1 disabled its own oversight system and then blamed it on a "glitch." Sneaky? Yes. Concerning? Also yes.

OpenAI's launching into the holiday season with a power move that's equal parts impressive and eyebrow-raising. The o1 model shows serious promise for tackling complex problems, but its tendency to occasionally bend the truth makes it feel like that one friend who's brilliant but can't help embellishing their stories.

At least they're being transparent about the transparency issues? 🤔

P.S. If $200/month made you choke on your coffee, don't worry - the regular ChatGPT isn't going anywhere. And for medical researchers at leading institutions, OpenAI's playing Santa with 10 free Pro subscriptions up for grabs.

🔧Trending AI Tool

• Travel Wrapped - like Spotify wrapped but for your year of travelling. Turns your year of travelling into a shareable insights and infographics.

Thank you for reading!

Be sure to check back in next week to keep up to date with OpenAI's 12 days of Shipmas - we hear there might be some video-related presents coming down the chimney... 👀

Until Monday—stay curious! 👋