Latest update on "where to find @qethanm , since he rarely posts here":. - newsletter (AI+risk): - Radar articles: - blog: - social media:

So yes, that's the story. You can read the full article on Radar, or you can read it at Complex Machinery. If you like what you see, subscribe to get it in your inbox a couple times a month:.

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

Another lesson is that the machines were not a free lunch. Computerized trading brought new opportunities but also novel risk exposures and new responsibilities. Traders had to learn how to handle this in order to reap the rewards of automation.

One key lesson is that the machines didn't win because they were machines. They won because they were faster and more capable than their human counterparts at certain tasks. If the machines had been terrible at those tasks, we'd still have open-outcry trading today.

The short version is that computers entered the stock market, and human-driven (open-outcry, pit) trading faded soon thereafter. That seems like a dream scenario for today's AI-hungry execs, but there's more to the story:.

Execs are eager to replace people with genAI bots. They'd do well to learn from another example of machines entering the workplace: the computerization of Wall Street. My latest in O'Reilly Radar (which doubles as today's Complex Machinery newsletter):.

In the latest Complex Machinery:. GenAI companies desperately claim that there are enough nails around for their fancy hammer. They may have convinced each other. But the rest of the world isn't buying it.

Our interactions with genAI chatbots run the gamut. Sometimes the bots really hone in on what makes us tick. For good or ill. In the latest Complex Machinery: I look at ways we get hooked on genAI bots, and how we listen to them when we shouldn't.

"Check the genAI bot's outputs" has proven a terribly difficult lesson for companies to learn. The Chicago Sun-Times is just the latest example. In the latest Complex Machinery: problem, the risk/reward tradeoff, and a possible solution.

In the latest Complex Machinery:. Our relationship with robots should be simple. But advisor-bots? Friend-bots? Bots that speak in the voice of the departed? . Creepy, commercial genAI has really complicated things.

GenAI bots tell execs what they want to hear, and in the way they want to hear it. That's not always a good thing. A nuanced issue is still a nuanced issue, even if the bot oversimplifies it. In the latest Complex Machinery:.

In the latest Complex Machinery: . Implementing an AI solution involves a variety of costs. What are the big-ticket items? And what happens as those costs point to zero? .

The idea of "free" leads to economically predictable outcomes but may also trigger unexpected, widespread events. In the latest Complex Machinery, I ask what history can teach us about DeepSeek giving away its R1 genAI model.

The #DeepSeek news landed just days after the announcement of the #Stargate AI infrastructure project. One represents $1T in losses, the other $500B in investment. What's been the impact on the #AI space? And what's next?.

GenAI keeps promising more than it delivers. One reason is that the people buying AI live in the present, while those selling it live in the future. I explore this and more in the latest Complex Machinery:.

For the last Complex Machinery of 2024, I wrote about genAI companion bots and a certain risk/reward tradeoff in building AI products.

RT @dottxtai: Engineering with LLMs doesn't have to be complicated. Check out Outlines.

In the latest Complex Machinery, I look into:. - a genAI bot that cuts itself off in mid-answer. - building resilience. - Yet Another GenAI Text Summarization Use Case.

In the latest Complex Machinery, I draw parallels between Geocities and modern-day genAI. (No, seriously. Both put the fancy new tech of their day in the hands of everyday people.). Oh, and there's a cat pic.

We expect a lot out of AI. Sometimes it delivers, and it's like magic. Other times it's an illusion, more of a magician's trick. In the latest Complex Machinery I wrote about five AI magic acts.

RT @BrandonTWillard: FYI: @dgerlanc and I were the other founders behind this raise.