Before copilot what I'd do is diagnose and identify the feature that resembles the one that I'm about to build, and then I'd copy the files over before I start tweaking.
Boilerplate generation was never, ever the bottleneck.
I can't even begin to imagine how a 12-year old who discovered how empowering bending the machine to do your will through code feels when, over time, realize that their dream career has been reduced to being an LLM middleman.
I always assumed that general advice targets the average person, without accounting for potential mental health problems. I cannot imagine myself trying to tailor an advice for all neurotypical people out there, because everyone can only share their own experience with the world.
>Turns out just running on autopilot most of the time is the healthier human experience.
I cannot believe that you'd argue for mindful nuance only to end up in such a blatantly general statement that contradicts everything you advocated for. That's without even bothering to argue how much of the time is "most of the time".
I'm kind on the fence, but not with the article. It's true that there are engineers who lean more towards one or the other way. For example, since the author brings up the switch from Neovim to VS Code due to features, I do love using Neovim for my TypeScript and Golang needs. But, if I were to work with Java or C# I'm switching to IntelliJ or Rider.
I believe it's healthier to attain some kind of pure-leaning... centrism(?) if we were to present pure/impure as black/white choices. I find it easier to imagine someone who deeply cares about squeezing performance through min-maxing to suddenly shift gears and deliberately introduce debt and just ship-ship-ship for the sake of pushing the product out, because they know exactly the price of the corner that they're cutting.
So I don't see it as a "you're either this or that" but more like "you should be this, and also be that when it's deemed appropriate".
Yep. I can see how relying on non-repetitive dialogue generated by LLM's will inevitably make us lose common ground scenarios that end up being memorable moments, like the arrow to the knee from Skyrim or "such devastation! this was not my intention" from Final Fantasy 14. A way to bypass this problem is to keep the important dialogue fixed, but this only one of the problems.
Another issue would be emphasizing the meaninglessness of the dialogue. For example, playing Trails in the Sky has lots of NPC dialogue that's repetitive, but at least the dialogue is relevant with how the NPC's life progresses in the grander scheme of things, such as having difficulty with her entrance exams, or having an argument with his fiancé. It's not main dialogue but adds flavor for anyone who cares about the world enough to interact with the citizens.
I don't think I'd like to interact with characters that I know whatever it is they have to say is generated on the fly and adds nothing other than random tidbits. The novelty would quickly wear off.
It doesn't need to be random tidbits. You can add "life events" for the character to the prompt, and steer the responses heavily toward the character wanting to talk about those. With an LLM in the picture, you can even have the character remember past interactions with you, and expand on it different ways depending on how you interacted with them.
A better application of LLMs could be merely using them as a text parser in text adventures. NPCs could continue to have some hidden, hard coded information or abilities, here stored in the context window, and the LLM is used to provide that information to the the player he (or she) puts in roughly the right words. Rather than exactly the right words.
Instead of generating truly random dialogue, you could choose to only generate random seeds within a certain number range, so that the probability that two players seeing the same line of dialogue increases, while still having enough variety for it to feel "random" to a solo user.
It's the whole debate about carefully created vs procedurely generated. LLM created dialog with be a vapid, vacuous and sterile as starfield was, or No Mans Sky
> will inevitably make us lose common ground scenarios that end up being memorable moments, like the arrow to the knee from Skyrim or "such devastation! this was not my intention" from Final Fantasy 14.
GLaDOS from Portal would offer one player pudding and another one a steak. You get to a wall which says “the ravioli is a fraud” and become utterly confused.
I agree. An important aspect would be lost. It would be as interesting and relatable as sharing that dream you had last night with your friends. “Uh huh. Cool story bro.”
I hate how true this is, even as a job seeker. I'm here building things using Next.js when I'm fully aware that I can build them using less than that, yet I must keep going and overcomplicate things in order to add my familiarity with specific technologies with my personal projects on my resume.
And then, there's people who do "resume-driven development" and push for more complexity in their workplace so that they can list real-life work experience for the next door to open. I know someone who made a tool that just installs Java JDK + IDE + DBearer using Rust, so that he can claim that he used Rust in the previous company he worked for.
I generally think we're more obsessed with being perceived as engineers than actually do engineering.
Just because you can never have absolute safety and security doesn't mean that you should be deliberately introduce more vulnerabilities in a system. It doesn't mtif we're talking about operating systems or the browser itself.
We shouldn't be sacrificing every trade-off indiscriminately out of fear of being left behind in the "AI world".
To make it clear, I am fully against these types of AI tools. At least for as long as we did not solve security issues that come with them. We are really good at shipping bullshit nobody asked for without acknowledging security concerns. Most people out there can not operate a computer. A lot of people still click on obvious scam links they've received per email. Humanity is far from being ready for more complexity and more security related issues.
The other grossly understated downside of lacking server browsers is how the player nowadays relies on the system to match him with the "best match" they can get. This opens the door to all sorts of skinner box manipulations, such as the game shoving you into teams where the probability of you winning is low, only to put you into a match where the probability of winning is high.
The ability to introduce randomization of reward around a layer of "skill issue" and plausible deniability for the matchmaker. Elo/bronze hell exists because even the worst players can just swing up and down their rank, whereas if you didn't had any other choice but play with whoever is in your local server (or LAN part, but I digress) then the only solution for you is to observe and adjust.
I'm from Greece and, we used to have lots of LAN arenas before fast internet connections became accessible. I'd get my face pushed by skilled people, and while I'd feel bad about it, the fact that I was playing with my friends and enjoyed myself made it all tolerable. Eventually I gave up feeling bad having negative k/d ratios, and could finally spectate and learn from others. The result was me becoming good enough to join my local CS clan. We never became best in the country, but I have really fond memories both from chilling as friends and highlights from matches.
My fear is that this will make the gap between newcomers and veterans so much bigger that the junior market will suffer more than ever before.
However, according to a few job openings in my area, "junior AI powered engineer" is actually a thing that some companies ask. Is it a good idea? I'd say no it's not. Do the managers who do the hiring while reading all the hype care? Definitely not the ones who ask for AI powered juniors.
You made me realize exactly why I love skill-based video games, and shun the gacha games (especially those with PvP). You swiped to gain power over players who don't. Yay?
The knowledge check will also slowly transfer towards the borders of fast iteration and not necessarily knowledge depth. The end goal is to make a commodity out of the myth of the 10x dev, and take more leverage away from the devs.
Boilerplate generation was never, ever the bottleneck.