On the bright side, most of us were never candidates for inventing relativity, really. I wonder if our mediocrity remains stable, of if we lose a proportional amount of capability as the luminaries did.
I think pens and pencils are mostly just preference and habit. I have some draftmatic mechanical pencils, nothing special really, but I’ve been using them for decades.
I suspect the real advantage of handwritten notes (for those who benefit from them) is that writing them fulfills a learned ritual for putting the brain in learning-mode. So, might as well match the environment as closely as possible, and prioritize familiarity over some quality.
Anyway, I can write obnoxiously small with my draftmatics, so I don’t see how the process could be optimized by a fancier pencil or pen anyway.
1) is there a reason both of the other responses to your comment are all full of Bro’s? It this an in-joke?
2) Regex is great, and vim is a good place to exercise the “try a regex” reflex. And on the regular old bash command line, it is great for making stuff like locate more precise.
You’d think that, but it sees like big business and governments are treating inference as somehow special. I dunno, maybe low temperatures can highlight this weird situation?
Temperature is an easy knob to twist, after all. Somebody (not me I’m too poor to pay the lawyers) should do a search and find where the crime starts.
Well, it's still not deterministic even at temp 0. The tech described in my comment's parent is speculative, and technically it's not even inference, once it's perfectly reproducible.
At that point it's retrieving results from a database.
EDIT: how would OP address my main point, which is that det. inference is functionally equivalent to any arbitrary keyed data storage/retrieval system?
> The tech described in my comment's parent is speculative, and technically it's not even inference, once it's perfectly reproducible.
This is not true. Fabrice Bellard's ts_zip [0] and ts_sms [1] uses a LLM to compress text. It beats stuff like .xz etc but of course is much slower. Now.. if it were non-deterministic, you would have trouble decompressing exactly into what it compressed. So, it uses a deterministic LLM
I guess, having just tried it, I use my pinky for []. But, it is not so bad, realistically the actions are
Hit [type whatever goes in here, hit ]
So it isn’t really a repetitive action or limited by the rate at which I can hit brackets I think even for programmers who use weird punctuation, a file should still mostly not be punctuation.
You aren’t scraping for the sake of training a model, but scraping the prices and availability is still scraping, right?
I think some of the folks running sites would rather have you go to the site and view the items “suggested based on your shopping history” (I consider these ads, the vendors might disagree), etc.
I’m more sympathetic to the people running sites than the LLM training scrapers, but these are two parties in a many-party game and neither one is perfectly aligned with users.
> scraping the prices and availability is still scraping
Web browsing is scraping, too.
I am not doing anything that I myself wouldn't do, it would just take me longer. I'm not mass-scraping, training new models, etc etc. I'm just using a helper tool to do some work for me.
If you prevent that, you are effectively saying: humans have to perform the manual labor of clicking and browsing through our site, they are not allowed to be helped in any way. I don't think this is the right answer.
If a human was being grilled like this by an LLM, I’d call that my dystopian. If companies have LLMs that address each other in a somewhat adversarial manner, that seems not so bad. They don’t have feelings to protect after all, so it is kind of nice if they can cut through each other’s bullshit.
Imagine if there were some kind of way to compress the interrogation down to known-valid aspects, avoiding the parts that are unnecessary for machines. You could have some kind of a programmatic interface...
Yea let’s call it the Agent Prioritized Interrogation interface.
Yeah, I take your point. It seems like the idea, though, is to work with services that are specifically trying to expose some kind of special LLM based interface. I dunno if that’s prominent or useful, I avoid that kind of thing.
Somehow these dumb displays always seem to be cheaper than the smart ones. For some mysterious reason all the chips and stuff to needed run an OS have a negative cost.
It is weird to think, computers are almost completely devoid of whirring nowadays. Other than the fan, and fans have gotten quite quiet. Floppies, CDs, hard drives. Tapes even (although I’m not that old).
It’s just kind of funny, I guess, the upcoming generation will never have the surprising “wow, my computer is silent” moment. I guess that was a one-time thing.
I tried to describe to my kid the sound of a 5.25" floppy disc the other day. MMWA MWA mmmmmm MWAMWA mmmmmmm.
He has seen 3.5" discs, but never the large floppies. His mind almost exploded when I talked about games needing 7 or 8 discs and hitting certain points where the game would pause while you put a new disc in.
I remember when the hard drive on my PowerMac 7200/90 started to fail (I was told that some of the ball bearings the platter rotated on were broken [0]). When the drive detected a 'wobble' on the platter, the entire computer would just power down with no warning at all - it was like the sound of a vacuum cleaner powering off. Silence and a blank screen that looked like a power cut, followed by that "oh sh!t" moment.
reply