What's your p(doom)? Plus Google deepfakes itself, new chips, and more!

After a brief hiatus due to a raging strep throat infection, I am now back, with all natural life inside of me currently being destroyed by antibiotics. But hey, my throat feels better! Long live medicine!

Today’s digest discusses a few major developments in the past few days. As usual, there is too much happening in Artificial Intelligence to cram it all in here. So, I am focusing on some of the most important things. 

Also, today, I am inserting a quick poll, which you can fill out below. Just trying to get an idea of my readers and what you are looking for!

And now, the digest!

First, A Meditation on AI Optimism and Pessimism

First, this observation was brought to you by our sponsor (me).

An article in today’s New York Times talked about how the term p(doom) 🌋 (“probability of doom”) is being used to categorize AI optimists and pessimists.

A lower p(doom) means you are an optimist, and a higher p(doom) means that you tend to fear that AI will lead to the destruction of the human species. People are apparently using this to sort out AI bloomers from AI doomers. The closer your p(doom) is to zero, the more optimistic you are. For example, if your p(doom) is 15, your vision of the future might involve having a picnic in a meadow with your robot best friend and flying home in an autopilot personal drone. If your p(doom) is over 50, you probably think robots are going to start killing us sometime after Super Bowl LX (or World Cup 2026, for you non-Americans).

I personally prefer using two p(doom)s. In my schema, the first p(doom) relates to humanity (not AI, which is part of humanity) destroying itself. It has a subscale in which AI occupies a small percentage (but is overshadowed by other human-induced cataclysms, such as nuclear/biological war, preventable disease, environmental destruction, etc.).

My other p(doom) is related to external events (meteors and other orbital bodies colliding with Earth, unstoppable natural plagues, the sun exploding unexpectedly, invasion of Earth by hostile aliens, etc.).

I don't really think about numbers for either one, but at this point, I would say my human-caused p(doom) is relatively high (above 50%) and inspires me to live in the mountains, generally far from people, significant military targets, and civilization in general.

My p(doom) for external events destroying humanity is 100%. However, if/when such a cataclysm occurs, there isn't much we can do, so what's the point of worrying about it.

Google Fake-Demos Gemini, Stock Value Jumps by $86 Billion

As I mentioned before, Google has released Gemini, its “all-powerful,” multi-modal AI (can read text, image, video). Their demo video has gone viral, with many viewers questioning how staged it is. One thing is clear: I would lose my mind if I had Gemini following me around and commenting on everything I did. 

Now, some media outlets have called out Google for “faking” the demo to make Gemini look good. Leading tech blog TechCrunch claims that the small print admits that the video has been manipulated to make Gemini seem smoother and faster than it really is. In the demo, viewers can see Gemini identify drawings in real-time, correctly guesses where the object is in a shell game (three cups), and understands that a person is trying to play rock-paper-scissors.

Despite this, these revelations take some of the shine off Gemini. It also makes the announcement seem a bit like a desperate attempt to get some attention. That being said, it did lead to Google’s stock price jumping 5% at closing on Thursday. Considering Google’s market cap of $1.75 trillion, that means the company’s value increased by $86.25 billion in just one day. Funny how that works, isn’t it?

AMD Sends Serious Signals to Nvidia

California-based chip, CPU, and GPU manufacturer AMD made a massive announcement yesterday. They claim that their upcoming MI300 chip will go into production in 2024, and that it will beat the industry-leading H100 by Nvidia. 

So far, Nvidia has dominated the AI industry with its chips, as most major AI frameworks are designed for their architecture. However, AMD has been a surging producer of chips and GPUs, rivaling established makers such as Intel, and already has agreements in place with Microsoft and Meta to supply their data centers with the new chips. Unsurprisingly, AMD stock rose over 10% upon release of the news. Noticing a pattern here?

New Tools, Gadgets, AI Toys, and Assorted Stuff

• Meta revealed its new, free image generator, Imagine.

• Freepik released Pikaso, a real-time AI renderer that transforms your drawings in real-time. 

• Researchers from Stanford and USC published a conceptual framework, Agent-Driver, to use LLMs in autonomous cars, improving decision-making and safety by making it more human-like. Hopefully, they don’t use people from the New York-New Jersey area for training.