ezst

21 ms·

511.
by ezst 1y ago
We have been in the phase of diminishing returns for years with LLMs now. There is no more data to train them on. The hallucinations are baked in at a fundamental level and they have no ability to emulate "reasoning" past what&#x2
512.
by ezst 1y ago
The point that is missed isn't about LLMs adequacy as a NLP technique, it's that they cost you 10000 times more for the same effect (after the upfront set-up), which is why I have my doubts that they will be used at scale, at the
513.
by ezst 1y ago
NLP was a thing decades before LLMs and deep learning. If one thing, LLMs are a crazy inefficient and costly way to get at it. I really doubt this has anything to do with scaling.
514.
by ezst 1y ago
Tech giants need to be dismantled.
515.
by ezst 1y ago
I mean, isn't it how it's always been, up until JS/python "single threaded" runtimes became what most people had ever encountered?
516.
by ezst 1y ago
I believe the point is less about "coloring" not having value as a type-system feature, and more about its bad ergonomics, and its viral nature in particular.
517.
by ezst 1y ago
That's so much less true nowadays, Web has become the default platform, where most people run most of their app/spend most of their time. Even Microsoft has had no choice but to embrace it, and Outlook (as in, the one from Micros
518.
by ezst 1y ago
> You can go and buy a framework laptop. > In fact, now is the best time in the last 20 years for either > having soldered RAM isn’t that big of a deal in reality to most people. Precisely. This whole sub-thread is a response to sa
519.
by ezst 1y ago
> Come on, there’s no way you wrote that down unironically and didn’t struggle breathing through the strong chemical copium smells. What a way to out yourself as some kind of irrational zealot. >> goes into barring the competition
520.
by ezst 1y ago
I suppose it's highly location-dependent. For instance, France recently banned smoking at the beach, parks and bus shelters.
521.
by ezst 1y ago
Depends on what you value. To me, MB Pro's keyboard is terrible, and MacOS is abysmal. On an ideological level, I defend right to repair, right to upgrade and oppose vendor lock-in. What does that leave me with? An admittedly decent CP
522.
by ezst 1y ago
As a non Apple user, yeah, M series are neat in the sense that the premium you pay goes into barring the competition from accessing the current nodes at TSMC, making Apple look good on benchmarks for 12-18 months or so. Apple used to have s
523.
by ezst 1y ago
And that's kind of an issue. Amazon effectively has a monopoly in this space, and competing at a similar level just is not possible anymore. And Amazon is so big that, even when you have a better product and service, it can buy you off
524.
by ezst 1y ago
> Still struggling to find anything in your wall of words that actually backs up your claims. I mean, one easy thing you could do is to explain how LLMs are adequate as >> thinking machines, search engines, knowledge archives, … Ho
525.
by ezst 1y ago
I think phabricator was doing a decent job at it while it lasted, don't know where they are at since IIRC it got abandoned and then forked. The best way to track meta history is to have it baked into the VCS, so here Mercurial is king,
526.
by ezst 1y ago
> We’re all operating off personal anecdotes and incomplete information. The difference is I acknowledge that. I'm sorry but I can't assume that your comment comes in good faith. You are the one making the exceptional claims he
527.
by ezst 1y ago
> The developer basically has to rewrite them and turn them into different apps to make them work in Wayland. Is that good or bad? Maybe if you hate a certain type of app it's good Just to make the record straight, it's not abo
528.
by ezst 1y ago
As I said, I don't reject LLMs for every situation (what you describe would probably fit the "iterative design" category). But as soon as this personal affair meets the wider world, inevitably becomes too big and transient to
529.
by ezst 1y ago
> there was a lot of hints of ChatGPT fluff. What does that mean? Are you implying that I used a LLM to formulate my comment? A comment about the dangers of LLMs for civic and respectful discourse? > Ultimately saying a LLM is only us
530.
by ezst 1y ago
I don't have a horse in this race, but I also can't blame the Wayland designers for not rushing to implement something like that. Encouraging application developers to manipulate windows themselves essentially means turning them i
531.
by ezst 1y ago
X11 was an unmaintainable pile of legacy mess, Wayland was its exhausted developers starting from a blank slate with tons of lessons learned the hard way. It's not a cabal or a conspiracy, just a sensible evolution.
532.
by ezst 1y ago
For me it's the AF. Every. Single. Time. People on smartphone, used to near-infinite depth of field, just forgot/never learned about focusing, and handing them a camera just too often results with the background being in focus and
533.
by ezst 1y ago
To me it's a useful reminder that those tools are nothing but text generating algorithms, optimised to produce compelling answers, irrespective whether they are truthful or not, having no concept of what's factual, and completely
534.
by ezst 1y ago
With evidence upon evidence that every centralised service eventually devolves into an enshitified user-hostile nightmare, why would you recommend yet another silo?
535.
by ezst 1y ago
> 1. I think folks would be naturally more skeptical of the government than they are of big tech, ideally leading to E2EE for email that's usable by the masses. That is so weird to me. "Institutions that exist for the sole purp
536.
by ezst 1y ago
It matters even more that Signal doesn't tolerate you using the client of your choice (like, one that doesn't push dark patterns and crypto in your face), or risk having your account suspended. It's time for people to wake up
537.
by ezst 1y ago
You are talking about productivity, I'm talking about knowledge. You may come-up with a product, then fire all engineers having built it. Then, what? It's not sustainable for a business to start from scratch every other year. Your
538.
by ezst 1y ago
Good of you to suppose that engineers cognitive decline doesn't translate into long term impactful business challenges as well. I mean, once you truly don't know your product and its capabilities any longer, what's left for y
539.
by ezst 1y ago
People praising VSCode's performance are probably better defined by having too fast computers than anything else, by all measures VSCode really isn't particularly lightweight nor performant.
540.
by ezst 1y ago
Add the usual army of bootlickers in the comments, defending those corporate interests against their own.

More ›