Everything You Need to Know About Google's New Pro-Tier Multimodal AI—Gemini Ultra—Released to to Rival ChatGPT

And, because I believe it's perhaps the most critical and unrealized area affecting society right now: what you need to know about "infoxication."

I often tell people to iterate with AI as they might a genius who is like 10 percent nuts. Sure, he might make things up occasionally. But hey man, he’s a genius. Gemini reminds me more of an irritating assistant who is constantly over-promising and under-delivering…but still, a very promising coder!

Hello all! This will be my last post for a little while. (Billing will be paused in a few hours after I hit publish on this so you won’t be charged until I write again.) Why: I’ve got a concrete offer to buy my site yougotthis.ai as an acqui-hire, and I need to prioritize and pivot accordingly. Once the sale happens I’ll be able to bring on some regular assistance to output this Substack (which I’m likely going to rebrand under the yougotthis.ai brand)—so then, even when a major upheaval happens it won’t just be me. (If you are yourself an AI or computer expert and are interested in working with me, please hit me up! Plan is to expand to podcast as well. It will be fun!) I appreciate your patience.

…So. Did you see all the AI commercials at the SuperBowl? Did you catch chip-maker NVIDIA pulling above Amazon in market cap yesterday? (No wonder: Nations are wanting their own sovereign AI capabilities.) It’s all incredibly overwhelming, right? We’ll talk about that in a bit, too.

But today’s topic is the historical widespread release (and rebranding from being called “Bard”) of Google’s Gemini AI.

• It beats GPT4 benchmarks!

• It’s a serious competitor to OpenAI!

• “…I put it in VSCode to see if it was better than Copilot. 10 minutes in I was literally swearing at it, and I don’t remember rage-uninstalling something so fast.”

Marketing Vs. Reality is striking right now in a major way amidst millions of users on message-boards everywhere who are sharing their actual (often frustrating) experiences with the new (rebranded) AI on the block.

Me? I will forever have a soft, squishy spot in my heart for Google AI.

See, months ago when it was still named Bard, I asked it about my beloved, named-by-multiple-outlets-as-a-book-of-the-year Simon & Schuster memoir Unwifeable, and it gave me the most incredible review in The New York Times I’ve ever read and informed me how it sold a whopping 100,000 copies in its very first week on release!

So totally cool, right?

Only problem? Neither of those things ever happened.

This right there is the major problem. Massive hallucination unlike I’ve ever seen. The kind that would inspire a nerdy, always vaping, subconsciously fame-thirsty engineer to think, “It’s alive!” Which, yeah, that happened. (Google Blake Lemoine.)

Testing it the past few days I’ve gotten lots of promises, lots of explanations of what it is working on, lots of questions back to me (nope, don’t want those) and lots and lots of confusion. But also some decent memes it created and in great news for programmers everywhere: okay code!

The main thing here is that this is Google. As unstoppable and undeniable as the returning current of the ocean waves. The fourth most valuable company in the world. Almost two trillion in market cap. So there is a lot more to know than just some snarky dismissal by someone who is spoiled rotten by the high degree of reliability of GPT4 via OpenAI.

Let’s go through all the key learnings I’ve discovered so you too can approach this new moment in the AI wars firmly engaged in that most wonderful of places that is located directly opposite of Hallucination Land: reality.