kolinko

14 ms·

271.
by kolinko 2y ago
LLMs, when used by an inexperienced developer will produce code that has a ton of fluff and is difficult to read/change. The worry is also that if devs don’t learn to write simple code, they will plateau very fast because they will not
272.
by kolinko 2y ago
Well it could be that the capacity is 3200-3349, and they publish the number that is in the center of confidence interval. Rounding up would be received badly, and rounding down suggests it has less capacity than it really does (on average)
273.
by kolinko 2y ago
Did you read any research on theory on mind and models? Since gpt4 they were tested using similar metrics to humans and it seems the bigger models “have” it
274.
by kolinko 2y ago
For most profitable search categories (ecommerce etc) I can see LLMs taking over - ChatGPT with Search is already more reliable than Google, not to mention Deep Research
275.
by kolinko 2y ago
With "death" flag set to false - sure. But are they collecting benefits / are they active in any way? Otherwise it's just that the death may not have been recorded. If Elon Musk wanted to be honest, he should've pub
276.
by kolinko 2y ago
I like planning as well (although many people don’t), but I still miss a ton of stuff that chatgpt’s Deep Research can find.
277.
by kolinko 2y ago
“A year ago” - wasn’t that gpt3?
278.
by kolinko 2y ago
Not a new model per se, but a new algorithm for inference - https://kolinko.github.io/effort/
279.
by kolinko 2y ago
I checked the original research by BBC but the underlying data was not published - no way to replicate the study, and no way to check its validity. They didn’t provide the answers models provided, so we can’t even say if the reviews of the
280.
by kolinko 2y ago
As for them not getting traction - many projects just don’t advertise which tech they use. Also, having an open source alternative - even if slightly worse - gives projects an alternative if OpenAI decides to cut them off or sth. What I tel
281.
by kolinko 2y ago
What specific piece of software are you missing on a mac? Aside from games, or industry-specific apps (like the goddamn KNX configurator, or CNC controllers), everything is there, at least for me, and with consistent and stable user experie
282.
by kolinko 2y ago
Proteins like reverse-transcriptase are quite complex, it’s not a mutation that would appear so suddenly. Also, having such a protein would significantly change how the virus operates, and it would be visible upon sequencing it.
283.
by kolinko 2y ago
Apple’s on device models are around 3B if I’m nit mistaken, and they developed some nice tech around them that they published, if I’m not mistaken - where they have just one model, but have switchable finetunings of that model so that it ca
284.
by kolinko 2y ago
Fun fact - in Polish we have separate forms for 1 (singular), 2-4 (plural but nit many) and everything else. Zero is in “everything else” 0 książek 0.5 książki 1 książka 2,3,4 książki 2.5 książki 5 and above książek 5.5 (any other fraction)
285.
by kolinko 2y ago
I also had doubts, but asked chat and it confirms it’s an issue - including sources. https://chatgpt.com/share/678b6b3e-9708-8009-bcad-8ba84a5145... The issue is that they are often localised, so even if it’s just 1% o
286.
by kolinko 2y ago
I have it similar - but when listening about math/physics. Usually play it to fall asleep. I think it's due to my adhd (diagnosed) brain getting the right amount of stimulation.
287.
by kolinko 2y ago
What? Central/eastern Europe than remembers communism definitely considers communists bad people.
288.
by kolinko 2y ago
With formal verification, what you want is correct statement of invariants / things to prove. As long as these are understandable, the readibility of proof doesn't matter much - it's the program job to verify that the proof i
289.
by kolinko 2y ago
One doesn’t exclude the other - a physical airtag may have an ID available, but not broadcast it anywhere. Also, “when there’s a will…” doesn’t really apply to cruptography
290.
by kolinko 2y ago
That's ad hominem. The only question is whether the research is correct, and the data is true. Having said that, with time we will have more independent studies.
291.
by kolinko 2y ago
Their Alpha* work from DeepMind is actually quite good and has a good track record. LLM/Gemini - yeah, what you said, I wouldn't trust a word their team says.
292.
by kolinko 2y ago
Define "our understanding". With complex / chaotic systems there sometimes are no higher level laws that govern them - all we have is just modeling and prediction.
293.
by kolinko 2y ago
Why have these features built in if you can have a usb back case/device with them? There are many back cases for specialised functionality like PoS, Ir camera etc. Also, with many cases you can have devices connecting wirelessly via bl
294.
by kolinko 2y ago
Care to share how you get advertisers and how you track how many listeners you have? I wonder how this compares to to the way online media works
295.
by kolinko 2y ago
It can be both.
296.
by kolinko 2y ago
Ah ok :) Still - I’d fund this game on kickstarter or wherever, to see it polished more.
297.
by kolinko 2y ago
Shameless plug of my dynamic algorithm: https://kolinko.github.io/effort/
298.
by kolinko 2y ago
The party game is meh, but the alien abduction is amazing. Managed to snatch all the people and it’s too bad there is no level 2. If it was a full game, I’d buy it on steam! Perhaps you could apply to the Steam Greenlight?
299.
by kolinko 2y ago
Not vitamins but vrious meds in my case. And a very chaotic lifestyle at times. The meds are always next to bed, and a pouch with meds for when I travel, plus a set of them in my backpack in case I sleep out of home. When I was in a relatio
300.
by kolinko 2y ago
With fire they count not only battery-initiated fires, but also fires caused by an accident. It would be nice if they compared it with regular ICE and other electric brand fires.

More ›