HNHacker News
TopNewBestAskShowJobs

jrm4

8,137 karma · joined March 19, 2020

submissionscomments
jrm4··on The Unbearable Cheapness of Open Weight Models
Technically correct, the worst kind of correct :)
jrm4··on I am a person who will look at the Steam Machine and cry
Yes, or and hear me out, go ahead and shell out the extra $200 or so as a "hassle fee," if you get the Steam Machine you're much more guaranteed that everything else will just work.

I absolutely agree on your notion of "what is with this 'I need the shiny new thing for sake of having a shiny new thing.' "

jrm4··on I am a person who will look at the Steam Machine and cry
Person has a Steam Deck that they don't use?

The only thing that keeps me from being genuinely baffled by this person taking the time to write this is the fact that I'm seeing similar takes elsewhere, which also baffle me to no end.

Are people really this enamored by "the thought of buying new thing," as opposed to, like, thinking about whether they'll use it?

Anyway, the Steam Machine seems like an extremely solid deal. It's a reasonably powered PC, and for an extra $200 or so, you'll get the guarantee that everything will just work, at least game-wise.

jrm4··on Steam Machine launches today
I feel like there's a midrange of "not particularly techy" gamers who will strongly appreciate - "I don't care about putting anything together, I just want to place PC games like a console."
jrm4··on Why Drawing Tablet Brands Won't Collaborate on Linux Floss Drivers
BRO THANK YOU.

This was something I knew to be true in my much more limited circle, but I very much appreciate the real life bigger example.

jrm4··on Why Drawing Tablet Brands Won't Collaborate on Linux Floss Drivers
I don't know why people keep making this point as if it matters; it sounds like you might be trying to absolve the creators or something.

But again, the people who gave the name to the project deliberately chose it because they found its slight offensiveness to be funny.

They knew what they were doing and chose to continue to do it anyway.

jrm4··on Why Drawing Tablet Brands Won't Collaborate on Linux Floss Drivers
I've been up and down this debate a million times, a lot of it here, suffice it to say -- the fact that you and others don't recognize this does not at all detract from my point.

To summarize, it's not e.g. about me being personally offended -- it's about people like me (a long time ago) wanting to show people this great software and other reasonable people seeing the name, understanding the meaning, and reasonably thinking "If this software were actually good, why does it have such a ridiculous and often offensive name?"

An unserious name -- literally chosen to be an edgy joke -- projects "unserious software."

jrm4··on Why Drawing Tablet Brands Won't Collaborate on Linux Floss Drivers
Preach. And it's a disease.

Signed, the guy who will forever believe GIMP could have been a contender with a name change decades ago.

jrm4··on There is minimal downside to switching to open models
At current "proprietary inference company behavior," investing in local inference sounds like the exceedingly far more rational option.

Long term predictability ought to far outweigh a few more cycles of performance.

jrm4··on There are no instances in ATProto
"Seems to have worked OK for email"

THANK YOU. It seems like far too few in this space really understand the benefits of actual decentralization.

ATProto feels like "centralized, except also we get other people to do the hard work with few of the benefits."

jrm4··on There are no instances in ATProto
Yes, and this is exactly why ATProto is worse and more dangerous. Instances are safer. precisely because they are more genuinely decentralized.

The ability to forever tie your stuff to a person, strongly, is exactly what the surveillance state would want.

Mastodon's model gives you plausible deniability. It's safer.

jrm4··on OpenAI Losses Increased Nearly 8X in 2025, with Spending Hitting $34B
Ha, not a problem.

Look, for coding and a lot of other things, AI is awesome.

But the here's the killer. I have a dinky 16gb VRAM card, and that's kind of the sweet spot for the level of AI I actually want. I don't want it doing too much, I'd rather create slowly than have it one shot something that I have to then pore over later.

Feels like a company investing kazillions in, i don't know, air-conditioning or building wi-fi. Yes, it's going to be around, and also no one's gonna need THAT MUCH.

jrm4··on Google Chrome update will close the door on ad blockers
Yup, this kind of criticism happens all the time. It's good to point out -- but a lot of people treat it like a deal-killer, which is the height of naivete.
jrm4··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
It should be readily apparent to anyone paying a little bit of attention that large corporations do not pay their fair share.
jrm4··on Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?
That was the point of the "e.g."
jrm4··on Google Chrome update will close the door on ad blockers
"Backed by tech companies"

Yup, huge red flag. Non-profit is the long-term way to go.

jrm4··on Google Chrome update will close the door on ad blockers
The complexity here is definitely a point against Brave.
jrm4··on Google Chrome update will close the door on ad blockers
This is an unadvisable forest-v-trees calculation. At any given time, Firefox is rarely the "best" at any snapshot metric.

It's nonetheless the thing you can overwhelmingly trust the most in the long run.

jrm4··on Google Chrome update will close the door on ad blockers
Brave seems silly theoretically for two reasons:

1) I think it's well established (or perhaps not well established enough) that "non-profit" is the only way to go for base level computing things like this. Profit motive (as distinguishable from "keep the lights on money," e.g. with Wikipedia) makes you do unnecessary and often harmful experimentation.

2) It's a fork of the thing you're trying to beat. Now, that may not be a deal-killer, but given Chrome's dominance, getting outside of that entirely seems smarter.

jrm4··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
Oh, so you are either deliberately, or ignorantly, spreading utter falsehoods. I think we're done here.
jrm4··on Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?
? I was speaking historically, not now
jrm4··on Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?
Having, e.g. seen Microsoft maintain a monopoly for well over a decade, there's nothing in my experience that suggests that "quality always beats hype" is remotely true.

It's entirely possible Claude is just winning the hype game.

jrm4··on Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?
But you're pretty much measuring opportunity cost in tokens per second, no?

I think it strongly remains to be seen whether e.g. tokens per second (multiplied or whatever by percieved quality of private model) actually means "better or more useful output."

I strongly suspect it does not. (though I also strongly suspect this will be very difficult to measure because the incentive to lie about metrics here will be so strong.)

jrm4··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
Not surprising.

The point of my first sentence is; private individuals and small businesses generally pay their fair share. Larger corporations emphatically do not.

jrm4··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
Yes, when they do so proportional to what they take, especially as compared to individuals and their tax liabilities.

You'll have to let me know when that finally happens, because that ain't now.

jrm4··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
What part of that said "defense?"

They can both be bad.

jrm4··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
“Well, Steve (Jobs), I think it’s more like we both had this rich neighbor named Xerox, and I broke into his house to steal the TV set, but I found out that you had already stolen it.”

-- Bill Gates

jrm4··on Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
Unlike the big companies who do this, which often are merely impure scams on tax payer money a little more downstream.
jrm4··on How to Earn a Billion Dollars
Percievable goods and services, then.
jrm4··on How to earn a billion dollars
But he seems wildly oblivious to the fact that there even is an argument to be had here, which there emphatically is.

Reasonable people can disagree as to the nature/extent -- or even the existence of -- exploitation, but this guy absolutely has impermissibly strong blinders on, rendering most of this article a waste.

← PreviousPage 7 of 34Next →