dexwiz

Tea party of the early 2010s also played by this rule. Lots of "I cannot recall"s which translates to "I won't even acknowledge wrong doing so why would I apologize?"

This is really helpful. During COVID I was looking into game engine design and the literature definitely confuses these topics. I imagine it's because everyone is writing about their specific experience but giving it the same name. It reminds me of MVC, MVVC, and MVP all being bundled under MVC.

How do you define old? The oldest initial release or the last update?

Yeah it was rhetorical. Search would be implemented as a tool call. Pure intelligence tests would likely have limited tools. But maybe they would have a python sandbox to solve issues like Rs in strawberry.

Is search part of the model or the harness?

Helps that Disney World was built up and not down. Everything is above sea level because it's in a swamp in Florida. Ground floor for guests is several stories up. I feel like Disney World is more comparable to an open air mall than a city.

Profit in video games as an industry is largely the gambling industry hiding under the another umbrella. I'm not even talking about loot boxes. Just old school gambling games like roulette and slot machines but in a digital format. Once you strip that away and loot boxes, the numbers are great for classic games.

Pandemic really ramped it up. Also screen addiction has gone mainstream. In front of a screen 12+ hours a day 20 years ago? Total nerd. Now? Standard behavior.

It's not the same, and there is no going back. We moved another layer up in the abstraction stack, but this shift feels fundamentally different.

I wonder if the people who enjoy the purely agentic version would also enjoy management. Personally, I have said no to more money because I enjoy the ground level work. But that work is just not the same.

I think running a model unsupervised in anything above medium is a sucker move to burn more tokens. The high effort models can be great in limited context, but unsupervised they too often end up navel gazing. High doesn't always mean smarter, but it always burns more tokens. Medium and low seem to be decent for day to day tasks.

Does anyone else hate reading AI summaries of code? Code can be pithy, but at least its terse compared to prose. When you add how verbose LLMs can be, I often end up reading a paragraph to explain a few lines. Or the opposite happens where the summary skips important edge cases or criteria. "You're right, X also does Y. I missed that in my initial analysis," is much too common of a phrase.

I like the idea of using LLMs to transform code into something more readable, and vice versa. I am not sure if meandering paragraphs and linear lists are the best targets.

How can you sanely project the last 12 months forward? We have seen a huge uptick in usage. Last summer AI was a toy to most devs, now every enterprise developer I talked to uses it every day. Coding agent providers are surely going to hit market saturation in the near future.

I think a better term would be bridle, but most people don't know it. But they do put harnesses on their kids and pets.

Anyone I talked to who actually built software found them to be bunk. Occasionally they work, but mostly they just burn tokens.

The only people I know still hyped on this approach were managers who were detached from actual engineering, primarily producing bespoke dashboard and tools.

You missed the crypto hype by a few years. Also zit is slang for a pimple or acne. Not a great name.