gibbitz

Exactly. SpaceX saves 70k by waiting for the chef to go bankrupt waiting for his day in court. It's the Trump playbook and is money and power weaponizing the court against justice.

Are you applying for a dev role or a PM/Account manager role? This feels like packing a parachute for a road trip or dropping a reference to Buñuel in a Netflix Adam Sandler film. Who's the audience?

Because the AI is a cyclical prediction machine that is more akin to a simple mechanism in nature like cell reproduction than a conscious dog. What has happened here is more like the AI company put bacteria in a petri dish and only put the lid on without pushing it down and is reporting the bacteria as a super bacteria because it contaminated all the samples in the fridge.

An interesting observation I've seen is that the AI companies reporting these incidents have largely been backed by venture capital where the self funded companies developing models have largely not had these "problems".

And for some reason, we don't do the same with software...

Someone needs to prompt "how do I letsencrypt?"...

If the point of the agent is to perform "reasoning" for operators, being able to understand the "train of thought" is part of the output. Removing access to this (efficient for the model or not) is removing value for the customer. AI shrinkflation.

I'm using it to code under free use on Cursor for the remainder of the month and finding I have to look over its shoulder to make sure it is following my spec and my prompt. So like... what you would expect it to do.

With banking I may as well do it myself if I'm going to use Grok.

Other prompts. Usually working 3 tickets at a time in parallel. My company's contracts make #tokenmaxximg a thing still. When this ends, I'll probably be researching another ticket to reduce context and token cost on my next prompt. More like making sausage than code, but that's the world we've always been living in as far as the C-suite is concerned.

Also the ash doesn't catch fire underground

I think the problem here is that we have a bias for those who agree with us that outweighs our patience in searching for oppositional opinions that are well founded. From a liberal bubble it is quite hard to find conservative viewpoints that aren't based on unobserved apriori belief or strawman arguments. It feels like being a heavyweight surrounded by lightweights and it makes me feel like I need a better sparring partner, I just can't find one...

IDK about the 2400 year old argument (apparently against paper), but if you need to use AI to organize your book find something else to do. If you don't have the patience or motivation to learn and do this, why should a reader have the patience to read the book. Most people who use AI in their creative process don't understand the creative process. Having something trained to do what came before to help you do something new is like putting guardrails on yourself to prevent you from doing something new. It feels right when you're doing it, but that's why they can charge for it. Like Coka-Cola -- It tastes great while you're drinking it so you buy it. After a while of this it puts your pancreas under distress leads to a myriad of health problems (diabetes, heart disease, kidney failure, cancer) that could have been prevented by just not drinking Coke.

Tell them instead that every time they use GPT for a paper they're killing education.

Even as a developer who uses RPI skills with models, I watch them write utilities that already exist, make inline functions when a utility class that it would logically fit in goes untouched. Bloat isn't just because of vibe coders, it's the nature of the technology. Writing code by hand is like drawing with a pen. The developer plans ahead, gets the shape of the code in his mind and writes just what is needed. A process of genesis. When using an LLM this same coder architects the shape of the code and asks the LLM to carve it from all the examples in its training data. A reductive process. The problem is that sometimes the shape doesn't confirm to the training data or the model doesn't share the developers architectural aesthetic. At this point imposter syndrome can set in and the developer accepts what the LLM wrote without the careful planning and thhought that went into the original approach. This leads to missed edges and curlicues in the code where an approach may not be fully understood (bloat). Since we're approaching the limit of quality that LLMs will reach, this will be what keeps software developers employed. Either building new systems or logically correcting brownfield codebases that have bloated over time. Having the patience to read through the bloated code and direct refactoring as well as being able to spot bloat/tech debt early will be the fundamental employable skills.

Meanwhile in the US the laws are being changed to protect the wealthy, while the current administration is siphoning value of national assets into their own pockets. The poorer the country, the more likely this is to happen. I wonder how long it will take for warlords to set up shop in Kansas...

In this case it's like cable TV. Only not a monopoly if you look at the whole country...