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.
1 combines with 2 because you can think about what changes to make to your code, and then that changed version runs. This is different from instructing an LLM to change something for you, where you have no insight in the underlying artifact being changed, or can’t predict what and how exactly the LLM will change it.
1 isn’t only about seeing the effects of your thoughts, it’s at least as much about transforming your thoughts into a more suitable, discrete and exact representation, whose effects you will in turn see. Having code as the intermediate representation that you work on, and that you work in terms of, is essential to both 1 and 2. Similar to how for a physicist, writing down formulas and working with them, transforming them, is essential.
The code is the language of thought, the model of thought, of what then gets reified as behavior at runtime.
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 ...
> 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
What instructions do you use to prevent the gruesome writing style? I have… not found this to be the case.
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 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 have not found this to be the case even with SOTA models as recently as last month.
It holds for a short while into the content but reverts very quickly to its default style.
The default style is so heavily embedded in the weights I don't think simply adding prompts will help.
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?
Whilst the objections you raise are valid, in the face of the country being destroyed if you don’t, the equation’s rather different. What does it matter if China is dominant if you’re better off living in the US?
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.
My interpretation of the argument is "Humans are also non-deterministic, but we make organizations work on a daily basis despite that. We could do the same with LLMs, which may in fact turn out to be less non-deterministic than humans."
I agree with both takes; organizations have evolved many error-correcting mechanisms like bureaucracy, laws and contracts to keep a gaggle of non-deterministic actors behaving deterministically. There is no reason those actors need to be human.
why do you think that something being non deterministic means it can't be better than humans at a task? hasn't this same approach been used when developing alpha go which has surpassed humans?
While the topic of this text is very interesting, it ultimately remains a bit unconvincing.
Nietzsche is depicted in an oversimplified way as a kind of tortured brute, and Otto Dix' late return to Christianity is rather exaggerated.
Dix remained interested in many topics until late in his life. While he radiates more empathy than Nietzsche, the one constant in Dix' art is that he never spared any feelings or aesthetic sensibilities. This urge to remain true to oneself and shed all masks is something he may have found reflected and elavated in Nietzsche's writing, even if he ultimately discovered that some of Nietzsche's admiration of fierceness was a peacetime naivety.
If you are just talking about contraception, that's a bit reductive.
There is a notable percentage of people who regret having kids, despite initially deciding to have them.
It is not possible to fully anticipate what it's like. For some parents, an unexpected change in circumstances may also change perspectives. E.g. their own parents may fall ill and add to the caretaking burden instead of sharing in it.
Ha! It's more than contraception availability although that certainly is why I qualified my bafflement as being in regards to reasonably intelligent people. It's pretty easy to imagine how this could happen due to temporary or permanent affliction by idiocy.
None of the personal examples I have in mind had circumstances that explain the resentment these men expressed. I think you're right that they simply failed to anticipate how parenthood would be, or maybe they didn't even recognize it as something worth consideration.
This is exactly it. I love my children more than anything, I would die for them. I do regret having children though, *at the moment*. I didn't realise just how very hard it would be. My eldest is autistic though, so for seven years we've doubted ourselves, struggling because of his extreme emotions and meltdowns. And then feeling judged by the entire world because I don't feel as the world prescribes.
Until you've lived that life, it is best not to judge. Just like I couldn't judge how I would personally react to being a parent until I became one.
Btw, my regret fluctuates, and there are periods where I would not change things at all. I hope, with time, those periods grow longer until they're the only periods.
Treat your kids as adults when they reach puberty, problem solved. Never act like they are too young for anything. Just explain how the world works and let them figure it out. Just don’t get mad at every stumble along the way. Be there for them if they need you.
I say this unironically. The “Move fast and break things” approach works wonders for the relationship between parents and children.
I guess from a pure language POV, one might argue that rust leaves humans with more opportunity to add abstractions that are too clever and too hard to wrap your head around. That doesn't feel like a strong argument, though.
Then there is of course the ecosystem, where go maybe has better libraries for some stuff (while rust may have better ones for other things).
> Go definitely has its share of problems for human authors because it's so verbose and boilerplate heavy, which means it's less of an issue with LLMs than it is for human coders.
If the premise of the article is true, and I think that it is, that's quite the downside for AI coding with go. The premise being that reviewing now plays much more of a role than writing.
Personally, I'd rather review, say, a ruby oneliner that extracts specific row values from a csv file with filter_map, compared to 40 or so lines of go, many of which I'd have to check individually for possible mistakes.
Indeed. My first thought was that this is a roundabout name for ordinary "rulebook slowdown", but of course there are other ways to increase the cost of useful behavior, not just rules.
It leads to an interesting way to think about company and civic health as well. Instead of focussing purely on incentives, one might assume that many are inclined to do good stuff anyway, and then ask: are we lowering the cost of all desired behavior as much as possible? And are we doing it for as many people as possible?
I also left yesterday. I thought the mood was still surprisingly calm, considering that there was already enough smoke in the city to make it look like a foggy winter day and smell heavily of campfire.
Now I am wondering if my return flight from Bordeaux in a few days will actually happen. Currently the airport is still operating, although the surrounding area has been evacuated. It really looks apocalyptic there, though:
Had to evacuate a few years ago: south-west of France. We could see 20 meters high burning flames on the closest hill. But weirdly enough it wasn't the most terrifying: the most terrifying was driving back (to the same place) at night time, in the summer, and feeling like it was snowing... for ashes from a huge fire about 20 kilometers away (?) were raining down the sky and covering everything. I could hardly see the road and didn't understand at first what was happening.
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.
reply