vivendi
@vivendi@programming.dev
- Comment on A great disturbance in the Fediverse 1 hour ago:
Saw Gererra would’ve thrown you to Bor Gullet for saying shit like this
- Comment on Fuck cars 18 hours ago:
A battery
A highly complex system with support systems to prevent it from literally blowing up (mmmmm tasty lithium)
A battery management system
A seriously complex piece of model specific code that is usually proprietary and has hardware level bricking fuses (so fuck you if it has a fault)
And also, all the other management shit like regen braking and what have you.
You’re doing a “reduction”. If you reduce literally anything to such high abstraction it seems simple. Yes, quantum physics is also " just " an observation of a continuous system, nothing to it mate
ICE cars can be completely reassembled with a couple of wrenches. The most proprietary thing in the whole system is the ECU, and you can swap that and whatever if you wish to do so.
- Comment on Cyberpunk 2 is now in preproduction, CD Projekt says 20 hours ago:
What? The cybernetic system is not a mechanic?
- Comment on Fuck cars 1 day ago:
90% of your power train is proprietary electronics that are designed to be opaque and break of you try and “tamper” with them
Most of an ICE car is tangible, rather widely replicable, blocks of metallic alloys
- Comment on Fuck cars 1 day ago:
So much easier UNTIL one of the six million proprietary silicon bullshit breaks (or planned obsolescence) and you’re SoL without official manufacturer support.
EVs are a maintainability and reparability nightmare, despite not needing maintenance as often
- Comment on Cyberpunk 2 is now in preproduction, CD Projekt says 1 day ago:
Go play Deus Ex in the same genre, or some cyberpunk based CRPGs. Those games lots of mechanics that play into the game, DXMD’s Augment system wipes it’s ass with Cyberpunk’s Chrome for one.
Also it would help if they didn’t rush the story so much, it was a product of crunch and it shows. You can fix the bugs but you can’t fix fundamental problems with pacing
- Comment on Cyberpunk 2 is now in preproduction, CD Projekt says 2 days ago:
There is a HUGE compilation on the subreddit for cyberpunk2077, but basically we got promised a vast, in depth RPG and instead got something mechanically on part with GTA Vice City and Call of Duty
- Comment on Cyberpunk 2 is now in preproduction, CD Projekt says 2 days ago:
2 FPS with Ngreedia AYEE AYYYY DLSUCKSS 81715X FRAME GENERATION enabled
The more you buy the more you save™ – God-Emperor Jensen Huang
- Comment on Cyberpunk 2 is now in preproduction, CD Projekt says 2 days ago:
It has a LOT of content cut from the board game, and the world is still incredibly shallow outside scripted gameplay
If you consider this anything more than a FPS game with cyberpunk aesthetics, I weep for your perception of quality
Fucking Watch Dogs 2 had a more interactive world, for crying out loud
- Comment on I hope i don't get downvoted for this 2 days ago:
Had a friend who was a freak ass like that actually
People’s dicks really are putting in overtime
- Comment on faen 2 days ago:
There is no way Norwegian is a real language
- Comment on ICE Taps into Nationwide AI-Enabled Camera Network, Data Shows 3 days ago:
My man, this is literally what they just did. This isn’t an strawman. Atleast google the meaning of your catchphrase ffs
- Comment on ICE Taps into Nationwide AI-Enabled Camera Network, Data Shows 3 days ago:
Something happens Americanly in America
Americans: “What are we, a bunch of üntermench asians???”
- Comment on faen 3 days ago:
Yes well google translate sucks
However datafag is rad as shit so I’m going to invoke law of cool vs boring
- Comment on faen 3 days ago:
You should bring back the usage of datafag as fast as possible
- Comment on faen 3 days ago:
Use it as a part of some other compound. It will translate fine.
For example, try slutt datafag lærd
- Comment on VCs are starting to partner with private equity to buy up call centers, accounting firms and other "mature companies" to replace their operations with AI 4 days ago:
But ERP is not a cool buzzword, hence it can fuck off we’re in 2025
- Comment on VCs are starting to partner with private equity to buy up call centers, accounting firms and other "mature companies" to replace their operations with AI 4 days ago:
You’re misunderstanding tool use, the LLM only queries something to be done then the actual system returns the result. You can also summarize the result or something but hallucinations in that workload are remarkably low (however without tuning they can drop important information from the response)
The place where it can hallucinate is generating steps for your natural language query, or the entry stage. That’s why you need to safeguard like your ass depends on it. (Which it does, if your boss is stupid enough)
- Comment on VCs are starting to partner with private equity to buy up call centers, accounting firms and other "mature companies" to replace their operations with AI 5 days ago:
The model ISN’T outputing the letters individually, binary models (as I mentioned) do not transformers.
The model output is more like Strawberry <S-T-R><A-W-B>
<S-T-R-A-W-B><E-R-R>
<S-T-R-A-W-B-E-R-R-Y>
Tokens can be a letter, part of a word, any single lexeme, any word, or even multiple words (“let be”)
- Comment on VCs are starting to partner with private equity to buy up call centers, accounting firms and other "mature companies" to replace their operations with AI 5 days ago:
No, this literally is the explanation. The model understands the concept of “Strawberry”, It can output from the model (and that itself is very complicated) in English as Strawberry, jn Persian as توت فرنگی and so on.
But the model does not understand how many Rs exist in Strawberry or how many ت exist in توت فرنگی
- Comment on VCs are starting to partner with private equity to buy up call centers, accounting firms and other "mature companies" to replace their operations with AI 5 days ago:
Broadcom management deserve gulag
- Comment on VCs are starting to partner with private equity to buy up call centers, accounting firms and other "mature companies" to replace their operations with AI 5 days ago:
For usage like that you’d wire an LLM into a tool use workflow with whatever accounting software you have. The LLM would make queries to the rigid, non-hallucinating accounting system.
I still don’t think it would be anywhere close to a good idea because you’d need a lot of safeguards and also fuck your accounting and you’ll have some unpleasant meetings with the local equivalent of the IRS.
- Comment on VCs are starting to partner with private equity to buy up call centers, accounting firms and other "mature companies" to replace their operations with AI 5 days ago:
This is because auto regressive LLMs work on high level “Tokens”. There are LLM experiments which can access byte information, to correctly answer such questions.
Also, they don’t want to support you omegalul do you really think call centers are hired to give a fuck about you? this is intentional
- Comment on Black Mirror AI 6 days ago:
I can’t really provide any further insight without finding the damn paper again (academia is cooked) but Inference is famously low-cost, this is basically “average user damage to the environment” comparison, so for example if a user chats with ChatGPT they gobble less water comparatively than downloading 4K porn (at least according to this particular paper)
As with any science, statistics are varied and to actually analyze this with rigor we’d need to sit down and really go down deep and hard on the data. Which is more than I intended when I made a passing comment lol
- Comment on Black Mirror AI 6 days ago:
According to arxiv.org/abs/2405.21015
The absolute most monstrous, energy guzzling model tested needed 10 MW of power to train.
Most models need less than that, and non-frontier models can even be trained on gaming hardware with comparatively little energy consumption.
That paper by the way says there is a 2.4x increase YoY for model training compute, BUT that paper doesn’t mention DeepSeek, which rocked the western AI world with comparatively little training cost (2.7 M GPU Hours in total)
- Comment on Black Mirror AI 6 days ago:
This particular graph is because a lot of people freaked out over “AI draining oceans” that’s why the original paper (I’ll look for it when I have time, I have a exam tomorrow. Fucking higher ed man) made this graph
- Comment on Black Mirror AI 6 days ago:
This is actually misleading in the other direction: ChatGPT is a particularly intensive model. You can run a GPT-4o class model on a consumer mid to high end GPU which would then use something in the ballpark of gaming in terms od environmental impact.
You can also run a cluster of 3090s or 4090s to train the model, which is what people do actually, in which case it’s still in the same range as gaming. (And more productive than 8 hours of WoW grind while chugging a warmed up Nutella glass as a drink).
Models like Google’s Gemma (NOT Gemini these are two completely different things) are insanely power efficient.
- Comment on Black Mirror AI 6 days ago:
China has that massive rate because it manufactures for the US, the US itself is a huge polluter for military and luxury NOT manufacturing
- Comment on Black Mirror AI 6 days ago:
- Comment on The Windows Subsystem for Linux is now open source. 1 week ago:
The infrastructure can be duplicated and studied tho. Would be cool if a little dreamy
