gunalx

9 ms·

151.
by gunalx 10mo ago
But they doubled the price g for this new mistralocr3 model to 2$
152.
by gunalx 10mo ago
I habe come across turning on caching means the llm has a faint memory of what was in the cache, even to unrelated queries. If this is the case its fully unreasonable to share the cache, because of possibility of information leakage.
153.
by gunalx 10mo ago
You can still tax based on the persons residence or citizenship. In the end someone can be attributed to wealth, and if they want to stay where they are physically, they should also tax like it.
154.
by gunalx 10mo ago
There was this one thinkpad that had a eink screen as its keyboard. Pretty cool, and you could use it as aregular display as well. dunno how linux would react though.
155.
by gunalx 10mo ago
Haiku 4.5 is a pretty decent small ish model. It conforms pretty good to my style guides, when cleaning up text, for eksample.
156.
by gunalx 10mo ago
It was at least, because of shitty security practices.
157.
by gunalx 10mo ago
API use was not merged in this way.
158.
by gunalx 11mo ago
Fair. Mostly the argument is, if all you need is to iterate on output to refine it, you get 10x the iterations, while lesser quality, its still a aspect to consider. But yes, why bother eine coding when they do make so many mistakes.
159.
by gunalx 11mo ago
I dunno. Even pretty weak models can be decently performant, and 9/10 the performance for 1/10 the price means 10x the output, and for a lot of stuff that quality difference dosent really matter. Considering even sota models are
160.
by gunalx 11mo ago
No those are (most provably) unpaid recommendations. Those are not the same as ads, because there is no economic insentive, and is strictly not ads.
161.
by gunalx 11mo ago
But why not just keep the consumer brand until stockpiles empty and blame supply issues until things possibly cool down, or people have forgotten the brand at all.
162.
by gunalx 11mo ago
You can go the waydroid style with namespacing, or native containers if using the linux kernel. No need to do a full vm
163.
by gunalx 11mo ago
I dont see the issue here. Its not like they have not disclosed what board it is based upon. And I do feel like its correct not advertising a mic if you dont have it enabled on this one. I dont really like nanokvm for being slow with update
164.
by gunalx 11mo ago
Have used gemini3 to GEW shot a few problems GPT5 struggled on.
165.
by gunalx 11mo ago
How did you get this to work? (See the config and did this before but never got ly to work on nixos, i migth have just been to early i guess) Ive been using LY on arch for a while, and it is great looking, when it works, but it has lots of
166.
by gunalx 11mo ago
Wouldnt be suprised if they said, we have about 2M users, lets estimate a common household per user, to inflate the numbers.
167.
by gunalx 11mo ago
You only really ned to deal with breaking api bianually though. I have really have had use of being able to quickly recover once my bootdisc fails and isntantly being able to have the same machine upp and running again.
168.
by gunalx 11mo ago
But that is not nessesarily enforceable in every region.
169.
by gunalx 11mo ago
Well problematic because they are using their own verifier as apanem of experts, with their own model trained specifically to satisfy this verifier. On the benchmark runs, they dont mention using human experts to cross validate their scores
170.
by gunalx 11mo ago
github pages?
171.
by gunalx 11mo ago
If i read it right it used multiple samples of itself to verify the aqccuracy, but isnt this problematic?
172.
by gunalx 11mo ago
Dosent just a whole lot of it boil down to agents being bad. Non deterministic. Dosent have the same references as humans. Cannot take simple clues. Stumbles on itself. Calling a agent junior level is a disserive to juniors. At least humans
173.
by gunalx 11mo ago
I was mostly ranting about open-webui and hoping onyx would be better than the current state. My usecase involves pdfs with lots of complex figures, ocrd through mistral ocr witch gives text, and images for figures (have tried multiple othe
174.
by gunalx 11mo ago
Yep, definetly this, They should have creds for open weigths, and bein transparent of it not being open source though. Pepole should stop being this confused when the messaging is pretty clear.
175.
by gunalx 11mo ago
Does it support multimodal documents? My main gripe with openwebui, in addition to it being slow is the fact that it mangles documents in the OCR step. tables that could have been understood great by an multi modal llm, just gets mangled by
176.
by gunalx 11mo ago
All of this can be done with ZigBee. It is open and local mesh over 2.4ghz.
177.
by gunalx 11mo ago
Not nessesarily. We have providers of hosted oss models (even large sota ones), with competitive API pricing. They have no reason to subsidice their services, outside og initial market grab. But there is so many of them that it dosent look
178.
by gunalx 11mo ago
Yep. Was using github for oauth on a petproject of mine. Got the unicorn, and was considering takingthe break, or just etting up something else. Seems to be running again for me now though.
179.
by gunalx 11mo ago
Isn't this from the model working o. really low res images, and then bein uppscalef afterwards?
180.
by gunalx 11mo ago
The mai reason I 2ill never buy any tplink consumer shit again, was that I could not setup the router without having direct internet access.(No way to SRT it up as lan only) And buying stuff that depends on internet services that shoudnt ne

More ›