davydm

I inverted most of my life, having started 3d gaming with a flight-sim-lile experience on Descent. I painstakingly switched to The Wrong Way on a recommendation from a friend, specifically to be more competitive in quake live. It took at least 2 weeks, but my scores show I'm better, even if marginally. And for games with no user rebinds, I can cope fine now instead of floundering about. It took a while to relearn using a controller this way too.

My initial stance was literally one of comfort, being used to things being that way. Researchers who claim that has nothing to do with it are just plain... wrong. I switched, on purpose, and took significant time to relearn control, mostly out of curiosity. I don't really prefer it this way, I'm just ok at doing it this way now.

No. Just please no. At best this is a hijacked Linux distro. At worst, it's an app no-one wants, marketing itself as an operating system.

I review code based on what's there. When I've reviewed ai-generated code, I always find issues. Usually subtle bugs that would just suck to track down. Sometimes glaring bugs that would break things immediately. Not once have I seen aigen code and thought highly of it. When I don't know it's aigen, I just assume that the coder is incompetent. Discovering that it's aigen, I'm colored completely unsurprised.

No, I don't want your ai slopcode. But if it was actually good, conformed to our standards, etc, I wouldn't even know. That's the crux - aigen code _is_ shit and that's why I'm biased against it.

And yet... Both are equally stupid. Using degrees of shitness to declare your wide the winner just means everyone lost.

have you tried rustdesk?

I've used teamviewer and anydesk, extensively, but have settled on rustdesk on my personal machine - works well, can run self-hosted or from the cloud.

top of the list, sindresorhus - watch out. Yes, he does opensource work. No, he doesn't play well with others.

I've had run-ins before, found actual bugs and he just stands firm on his pedestal.

it's at the point where if i see that he made the package, i try to avoid it - if there are issues, his ego is too large to allow community assistance.

Cool idea, but probably not that useful outside the US, where most people don't waste food like Americans do, at least, that's what's portrayed on tv and movies from America. We're always appalled at the giant spreads of rood which are summarily tossed into the garbage disposal or bins in American entertainment media. We teach our kids to take what they're actually going to eat, we don't produce 10 servings for 4 people, etc. There is no actual food in my garbage that goes out every week. When there are small scraps, that's doggy treats.

I suggest an open kanban approach:

We use Trello, but any simple cards-and-lanes service would suffice. You need at least 4 columns:

Backlog, blocked, in progress, done

You add cards to backlog, and self-prioritise. You provide access to the board to someone who has the say in your company, so they can reorder and comment and so on. You periodically review the order - perhaps something you just did makes it easier to accomplish one of the other cards, say. You always just pick the card off the top, put it in "in progress" and work on that. If you're pulled off because something else is of higher priority or you're waiting on some other person or process, move the card to "blocked", add a comment as to why. Pick the next off of "backlog" or "blocked" (if you can) to go into "in progress". When you're done with a task, move the card to "done". Perhaps set up auto-aexhive so that cards older than, say, 3 months in "done" are archived (still available, but not cluttering the board).

Sorry if I'm just explaining what you already know or do. This is just what I'd recommend for "eating the elephant": organise your bites and tuck in.

not working for me on Firefox Nightly, even though that has wasm support, it tells me "your browser does not support the WebAssembly (WASM-GC) version of this game".

I'll have a look in "that other browser", but it would be really nice if it worked everywhere (:

nice idea, but built on the faulty assumption that ai summaries are all the truth and nothing but the truth - if you're missing vital facts or if ai has hallucinated a bunch of bullshit in there and you're none-the-wiser, you're now actively dumber from ingesting it.

Perhaps this is why so many world leaders seem to be not-quite-right? Perhaps they are already leaning heavily on the gospel according to AI, chapter and verse, instead of actually understanding the material. Ironically, they're in the same position of knowledge as the AI: no understanding.

There is already at least one company successfully doing this: https://www.offerzen.com/

I'm not saying "don't compete" - this is a good thing for devs.

Crappy, bloated, slow, buggy, frustrating for customers. This is modern enterprise software. Older enterprise software was mostly just bloated and stagnant.

250k lines that are "carefully reviewed"? "Enterprise grade" I'll buy because enterprise code, especially now, is pretty shit - slow, buggy, insecure.

I guess we all wait for the output, which is the only true proof. Right now, it's a loon getting excited about llms competing with each other to produce what is quite likely way too much code, within which lurks way too many bugs. Any seasoned programmer will tell you less code is better, both for performance and for reliability.

I'd say I can't wait to see the output, but at this rate, it will likely be 1m loc when done, and if I was going to review a million lines of code, I'd be looking at the Linux kernel - a carefully crafted bit of software with actual review processes in place, that's used in millions upon millions of devices. I can pretty-much guarantee that whatever slop is finally served up, it won't be anywhere near the same league.

who still bothers with this after the t-shirt shenanigans and resultant hostility towards genuine contributors?

i was so excited for the first and second ones, and then people, true to form, ruined it

for now - and when people realise what a mistake it is to get the "bag of words" to write code, they'll be back - at least, some of them, as others will have moved on to survive.

FTA: "But with billions of dollars pouring into generative AI every year, we will see instant voice-to-code capabilities + bug-free quality in 2-5 years."

HA HA HA HA HA HA HA HA HA HA HA HA

omg, thanks for the laugh - "bug-free quality in 2-5 years" pfffffft I'm not holding my breath - rather, I think that by then, the hype will have finally lost some steam as companies crash and burn with their shitty, "almost working" codebases.

fta:

A - “true & useful, AND new to me”

B - “true & useful, but NOT novel”

U - “unclear, undefined, unknowable, or just flat out wrong”

This clearly defines my experiences with codegen ai - 1% A, 30-50% B (true, but not useful, like telling me the --count parameter sets the count) and the rest U - loads and loads of complete bullshit, sprinkled with enough confidence and subject-appropriate wordage that it looks valid, but someone who actually knows what they're doing will find the flaws - ironic, since it's people who _don't_ know what they're doing who are the major target audience

South Africa

it's not a panacea, but sticking to strict tdd also really helps, for the same reason too: it makes you think about the problem before just jumping in to "solve" it

if you're doing it strictly, you should be following the red/green/refactor approach: 1. write a test which inches forward the expectations of capabilities from the system - the smaller the step, the better; run this test and verify that it fails for the right reason(s) - ie, that it's actually failing at the point you need to extend 2. write only the code required to solve the test - this may seem obtuse at first, but even obtuse solutions (eg a function which just does `return 0;`) are a valid start - rein in your desire to solve the whole thing right now - no test, no code; now re-run the test until you get a green, tweaking if you have to to pass the test. Reaching for a debugger is valid, but shouldn't be your first choice. 3. refactor - this is the part a lot of people leave out. It doesn't just apply to the prod code either - is there a pattern emerging in tests? could you simplify the test to make it easier for a stranger (you in 2 months) to read and understand what's going on, what the test requires, and, if it breaks, why?

the process forces you to slow down a bit into steps of incremental extension of the system - it forces you to re-read your code at least twice, and to have a proper plan before writing production code. It provides a test suite to run against regularly to prove that regressions haven't been introduced. You should 100% have this running somewhere, and github actions have made this really trivial to have continuous-integration-style testing on your code.

yes, add more ai slop-crap, it's definitely what everyone needs rn

the companies who think they can replace juniors, who learn, adapt, and improve, with llms, are going to have a nasty shock - some already are, others are banking their shocks for later

my company hasn't hired juniors in years, and the ceo is adamant that he only hires seniors, but i think they're missing out - juniors offer new perspectives and opportunities for seniors to improve their social skills via mentoring

also, the number of times i've seen something like "help, I need a coder to finish this project I started with chatgpt - 16k lines, and it nearly works" - good luck, throw that garbage in the trash, where it belongs

Similarly, a lot of people start out thinking that money will solve their problems, and it does solve a lot of problems, but people get onto the treadmill of earning more and spending more, when happiness comes from contentment with what you have, not discontent and constant struggle to "get more stuff"

Not saying don't be ambitious, not saying don't try, etc - just saying that once you've reached a level of not having to worry about essentials (food, clothing, housing, water, etc) with enough "headroom" to also have fun in life, do you really need more? Will you be happier with more? Really happier? Or will you just get a short burst of happiness, followed by hollowness as you realise that no matter how much you earn, you can blow all of it on crap.

So, instead of being forced into a bubble, users can choose to be in a bubble of ignorance. Yep, that aught to solve it.

grifters keep on grifting

this is not the local ad-rag - gtfo

I'm shocked! Shocked, I tell you!

hypebro gotta hype

you first, my dude - if the ais are so good, you can prove it.

No.

Pull the wool back a little, peek outside, and see how the grifter techbros are still pushing for that bleak dystopia where they will feel validated by the masses who haven't hoarded billions in wealth.