"He's investing his cash into VC funds" is the tell for sarcasm - any founder would put that money into their own company, and a lot of founders take VC money even if they don't need it because the philosophy is "if they are offering, don't say no".

It's entirely possible that this is sarcasm, but it's getting harder to tell. What I can tell you is if this was true, you'd hear about it endlessly.

BS. Who was the company? What is the product?

No, I don't think about it because I'm fairly well adjusted psychologically and am not a hypochondriac.

It's like using GenAI to summarize notes of a meeting nobody attends or cares about.

How important could a hire possibly be if you are OK with a hallucinating bot running the interview?.

Anecdotally I am hearing the opposite - job sites flooded with AI generated applications, read by AI bots to screen and then bot assistants in the interviews.

After the moon landing in 1969, there were a lot of people predicting we would have moon bases within a decade, and people would live there. The idea made no sense, as the logistical problems of such a hostile environment outweigh the benefits.

Incidentally, the whole reading an MRI to detect cancer is image classification - and has been around for decades. The difference now is we have way more compute available much more cheaply, AND more digital data to train the algorithm.

The other thing that happened in tech was the "Great Resignation" - where people were quitting their job to make 30-40% more doing the exact same job. Well, businesses have realized they overpaid these people and the overall macro environment has seen declining demand.

In other words, this is just normal business cycle stuff - but e/acc wants to believe this is all due to AI replacing workers. There is no objective evidence for this.

Big Tech firms have been bloated after a 10+ year bull run thanks to 0% interest rates and massive pandemic era QE, which increased demand for digital/infrastructure. AI is the next leg of this capacity demand.

What *is* true is big tech is redirecting spend towards their AI data centre projects, and these layoffs could be part of that. Businesses don't axe productive divisions - there is obviously some other things going on.

Simple apps that don't do anything useful can be "vibe coded" because the stakes are low, and the cost of error is low. Complex, production applications with several integration points that need to be reliable fall apart quickly in this paradigm.

Anyone who has worked with code assistants can tell you they don't replace anything. They are useful for a narrow set of applications where there's a lot of boiler plate code in public repos used to train these models.

An obvious trope on X is that these layoffs in tech are because people are being replaced by AI. This is not correct.

Cool? Yes. Useful? Somewhat. Valuable? No. People would want such an app to be free.

I can think of tons of cool applications for this GenAI tech but they are low value. For example, when I am looking at a busy shelf in a grocery store, it would be cool if you could ask your phone where a certain product is and it could highlight it in the image.

In the majority of business workflows, the outputs are deterministic. They are based on a set of rules that can be encoded in a system, and the outputs need to be reliable. This is why attempts to use LLMs in real-life workflows has been failing.

RT @Dr_Gingerballs: The entire US economy is propped up by financing companies to give away compute.