alecbz

5 ms·

31.
by alecbz 6mo ago
> if enough people want to willing vote in a corrupt president Why do people do this though? Maybe it's inevitable, but I think there was a lot of pent up frustration with the government that led a lot of people to just say "fu
32.
by alecbz 6mo ago
> Although [touch typing] refers to typing without using the sense of sight to find the keys ... the term is often used to refer to a specific form of touch typing that involves placing the eight fingers in a horizontal row along the mi
33.
by alecbz 6mo ago
Say more.
34.
by alecbz 6mo ago
Craft, in coding or anything else, exists for a reason. It can bleed over into vain frivolity, but craft helps keep the quality of things high. Craft often inspires a quasi-religious adherence to fight the ever-present temptation to just cu
35.
by alecbz 6mo ago
A "deep" tool that fully automates fairly specific tasks works this way. LLMs are more of a "shallow", general tool that can partially help with lots of different things, but none so completely that they alleviate the ne
36.
by alecbz 6mo ago
If the 1% is just a bit less efficient with the new tech, sure, but it's different if the 1% means your car crashes.
37.
by alecbz 6mo ago
Sure -- I think articles like this are a warning that the skills we're losing are likely _not_ so completely supplanted by AI that they'll soon be irrelevant.
38.
by alecbz 6mo ago
A car that can self-drive 100% of the time is a new tool that could make driving an obsolete skill. A car that can self-drive successfully 99% of the time is dangerous because it trains people to not be ready to take over for the 1% they ne
39.
by alecbz 6mo ago
Big Tech is not the same thing as all technology.
40.
by alecbz 6mo ago
I'm sure some companies do this poorly but there's lots of places where code review happens on every PR and there's processes and systems in place to make sure it's an easy process (or at least, as easy as it should be).
41.
by alecbz 6mo ago
>sure, they coined the term “move fast and break things” Yeah I'm aware, but as any company gets larger and has more and more traffic (and money) dependent on their existing systems working, keeping those systems working becomes mor
42.
by alecbz 6mo ago
There's a lot of software in between Air Traffic Controller and Facebook. And honestly would Meta be okay with Instagram or Facebook going down even for just a few minutes? I'd think at this point that'd be considered a fairl
43.
by alecbz 6mo ago
Yeah in many places we had two humans with context on every line, and now we're advocating going to zero?
44.
by alecbz 6mo ago
How do you know which lines you need to review and which you don't? Does it feel archaic because LLMs are clearly producing output of a quality that doesn't require any review, or because having to review all the code LLMs produce
45.
by alecbz 6mo ago
Yeah performance optimization is potentially another good area for LLMs to shine, if you already have a sufficiently comprehensive test suite, because no functionality is changing. But if functionality is changing, you need to be in the l
46.
by alecbz 6mo ago
> The best experiences I have are those where I can describe what I want done with details. But that's the hard part! You can only eke out moderate productivity gains by automating the tedium of actually writing out the code, becaus
47.
by alecbz 6mo ago
Stuff like this works for things that can be verified programmatically (though I find LLMs still do occasionally ignore instructions like this), but ensuring correct functionality and sensible code organization are bigger challenges. There
48.
by alecbz 7mo ago
It's good at writing/updating tedious test cases and fixtures when you're directing it more closely. But yes, it's not as great at coming up with what to test in the first place.
49.
by alecbz 7mo ago
I'm still choosing to believe this is all a joke.
50.
by alecbz 7mo ago
If AI makes people more productive then labor is cheaper than it was pre-AI, even at pre-AI salaries, because you're getting more done at the same cost.
51.
by alecbz 7mo ago
All words in a thesaurus would generally also be in a dictionary? The difference between a thesaurus and a dictionary is what each tells you about a word.
52.
by alecbz 7mo ago
"to be" is a very weird example because that's just the full infinitive of "be" which is definitely in dictionaries: https://www.merriam-webster.com/dictionary/be
53.
by alecbz 7mo ago
I think 'phraseme' is closer: https://en.wikipedia.org/wiki/Phraseme
54.
by alecbz 8mo ago
It's not that they can't fix bugs at all, but I find that if I've already attempted to debug something and hit a wall, they're rarely able to help further.
55.
by alecbz 8mo ago
> "Seem". "Feel". That's the anthropomorphisation at work again. Those are descriptions of my thoughts. So no, not anthropomorphisation, unless you think I'm a bot. > These chatbots are called Large Lan
56.
by alecbz 8mo ago
I wasn't flying 25 years ago but I'm not sure what you mean, or how that's relevant actually. The point is just that it takes them more time to do the "extra screening" if you don't have your ID than the standa
57.
by alecbz 8mo ago
This happened to me once, they just brought out someone (supervisor?) who asked questions about what addresses I've lived at, other similar questions I'd probably only know the answer to. It does take longer than regular screening
58.
by alecbz 8mo ago
It's not that they'd pay individual employees more, it's that they'd hire more workers to account for the fact that their existing workers are tied up doing extra verification. Though they might not do that either.
59.
by alecbz 8mo ago
This feels too strong IMO. Just focusing on the outputs we can observe, LLMs clearly seem to be able to "think" correctly on some small problems that feel generalized from examples its been trained on (as opposed to pure regurgita
60.
by alecbz 8mo ago
Trying to get LLMs to understand bugs that I myself am stuck on has had an approximately 0% success rate for me. They're energetic "interns" that can churn out a lot of stuff fast but seem to struggle a lot with critical thin

More ›