jaynetics

I agree. Perhaps too many blogs go for a single, coherent punchline, and the text form of "somewhat scattered thoughts" is under-appreciated?

I have no idea if that's the case, but it would be funny if it didn't support negative prompting and they got someone to write the prompt who wasn't aware of that.

Edit: Okay, I read the prompts and they use negative phrasing a lot, so I guess it does work, but boy, are these prompts laborious. Requirements are repeated ad nauseam. This model seems to be quite tone deaf.

At least for us dev types.

It's not really contradictory, minor changes can have major effects.

> Acceptance (people who are focused on the problems that get worse as AI gets better - social disruption, x-risk, government/corporate control)

You can consider all of these problems in the other stages as well, even if you still doubt the impact on your own field. Most devs in the stage of acceptance, however, will be increasingly preoccupied with how to keep paying their bills instead of such lofty questions.

The word "testing" is a bit misleading. It's not about test coverage, it's about experimenting to find architecture decisions that are suitable for further development. So, not generating automated tests, but instead building throw-away features on top of the new thing and seeing if they turn out okay. (Though the post remains a bit vague on how to judge "okay".) Maybe it should be called something like "ephemeral build-out" or "future usage trial" or so.

I think your argument has two facettes:

1. It is great to see an effect of your thoughts, like something happening in the real world. 2. Coding offers a feeling of perfect control due to the deterministic nature of code.

I would argue that 1 is still the case with LLMs, and 2 was always just 95% true, never 100%, because of oversights, bugs in libraries, and what have you.

So what actually decays is maybe not so much the feeling of self-efficacy, but the feeling of control, and even this is a gradual change. Perhaps it would be less unsettling if it didn't happen at the same time as a potential loss of control in other areas: career, salary, mode of working etc.

> It holds for a short while into the content but reverts very quickly to its default style.

You can always make it do a second pass, and if that isn't enough, even let it fix each generated text block individually.

Currently I append the snippet below to prompts where I care about the resulting text. Just putting it in the context up front isn't as effective, and I assume if the overall prompt is large, or the context window is full, it will also be less effective.

> Use plain, clear, everyday language with a linear deductive flow; avoid hyperbole, juxtapositions, metaphors, analogies, punchlines.

I think most developers and most dev work is now in the field of websites, apps, and other stuff that is by nature connected to the internet, and has some relatively monolithic backend that has not been built with "blast doors" or other internal safeguards. I guess it will take considerable time until most software is in a state where the impact of AI changes can be reliably estimated without looking at the changes themselves, and some patterns for how to achieve this probably still need to be discovered.

> maybe in the future we need to all be more concious about leaving LLM-tells out of our writing.

I expect the gruesome writing style of LLMs to be fixed relatively soon, seeing how it can already be prevented with just a few lines of stylistic instructions.

I just wonder: how long will it take us then, after the fix, to not be thrown off every time we see some unnecessary catchy or contrastive phrasing?

Being traumatized by poor writing was not on my bingo card for 21st century technological progress ...

So, do you have a counter-argument? Anything that humans can do that goes beyond "pushing probabilities higher"?

I like the thought, but it seems daunting. What about global competition? If some other countries have few or no such taxes, how do you deal with that? Do you outlaw the outsourcing of all white collar work? Outlaw all remote work as well? Will you have to suppress a local black market for intellectual labor, built on self-hosted open source models?

Humans are also non-deterministic, for all practical purposes, and the more of them you need to build some piece of software, the harder it can get to keep the likelihood of issues low.

I'm not saying it will happen, but at least it's not inconceivable that LLMs could surpass them in this calculation more and more often.