@MegaBasedChad Get real drunk and review the worst code you've ever seen in your life
English
s
8.6K posts
s
@idoccor
I have made entire HFT systems from nothing
Brooklyn Katılım Ekim 2012
1.3K Takip Edilen4.1K Takipçiler
It’s called the Ballmer Stack and it’s how you stand out when the whole city is on stimulants
English
So quad witch opex and end of war bring us to 8k SPX? @DeepDishEnjoyer may not mentally survive it gang
English
@owainkenway I just didn't have time for that. There was a grad student looking at porn in the office that I had to report though.
English
Back when I started out being a sysadmin, the sysadmin in charge gave me the talk which is basically: “the reason you don’t look at the users’ files is that if you do find something illegal you will have to call the police” which has killed any urge to snoop.
Alek Asaduryan@Ldnbox
SaaS founder realizes they may have accidentally built the backbone for a drug trafficking operation. 😜
English
@ScottPh77711570 Yeah I'll say, he follows 10k people and I'm not one of them. What a dickhead
English
@davepl1968 The tradition continues because the STL in msvc is still scary to this very day
English
I had this conversation at Microsoft in 1996:
Me: "Why do we have our own pointer array code?"
Mgr: "Because it's solid and well tested."
Me: "So is vector<> in the STL!"
Mgr: "Devs don't know the STL"
Me: "They're devs, they should know the STL!"
Mgr: "That's great, but they don't, so no."
And so we continued to use and write all of our own containers and so on. Because the STL was scary.
trish@_trish_xD
i used to roll my eyes whenever senior devs said "just use the standard library." i was wrong. they were right. so much third-party stuff is genuinely unnecessary.
English
@HikariTaiyo_ch Yeah never trust the memory card. I still have never seen beyond about 2/3rds of ChronoCross
English
My dad took 2 weeks leave from the Marines to sit down on the couch, turn this game on, and I swear to god, I saw it with my own two eyes, that PlayStation did NOT turn off for 14 days.
１６ビット📀@favoga_
当時リアルタイムでプレイしてた人に聞きたい。FFⅦって発売当初からすげぇぇてなってたの？
English
@maxwell_marlow A good way to incentivize raising the minimum wage probably would be to make it the pension pay rate though.
English
This would cost about £570bn, btw.
Looney bin economics at play, but obviously the only consideration is Me, Me, and of course, Me.
Charlie 🏳️🌈@burnleyboi97
It must be fun to be have boomer levels of entitlement
English
s retweetledi
@grugcapital Nothing gives me meaning in life like finding out what Robert DeNiro think about politics
English
Guy who listens to Passion Pit and thinks “God I wonder what their opinion on global warming is?”
You’re mentally ill
alco ⊢ ꙮ@qualiascript
being right-leaning and high openness is so funny. "this is one of my favorite musicians, i disagree with everything they stand for, highly recommend"
English
AMD users i need your data. i have almost zero AMD configs in my benchmark collection and i can't answer these questions properly without real numbers.
if you're running any model on Vulkan or ROCm drop your GPU, model, tok/s, and engine below. 7900 XTX, R9700, 6800XT, anything. one reply from you fills a gap i can't fill from NVIDIA alone.
Just some person@ThatWolfieGuy
@sudoingX AMD AI Pro R9700x2?
English
@LottoLabs q3 120b-a10b is about the same performance as 27b and works about the same as q4 as far as I can tell. It's my favourite so far
English
Qwen 3.5 models ranked on 3090 W/ hermes agent.
0.8b: for fun, cpu usage, don’t expect much but it runs on anything
2b: starting to be usable, can do small tool calls (not super reliably), drifts from tasks easily, major steering required
4b: actually usable, follows tool calls reliably, follows skills reliably (major bonus), doesn’t drift from tasks as bad as 2b.
9b: all of 4b but more capable w/ more complex tasks, still needs steering, still not 1 shoting tasks but more intelligent than the smaller models
A3b: fast, more general intelligence, feels like the 9b speed but the reasoning closer to 27b, follows tool calls and complex skills well, minimal drift, just lacks big model coding abilities.
27b: the 3090 goat imo, no drift, tool calls for days, writes and follows skills very well, feels like sonnet 3.6-4 range of knowledge level with less glazing, code is usable and can deal w/ multiple files in projects. General knowledge level just feels higher. Only downside is it is slower than A3b and 9b obviously.
English
@kenwheeler Is that the one where Ben Affleck was involved in the production and somehow comes out at the end winning and looking cool as fuck?
English
watching the town makes me wanna move to boston and ruin my life
English
Here's Sandfall Entertainment using the same techstack (Metahuman) but their scanning equipment is only an iPhone.
Turns out if you know what you're doing and have willingness to learn, it's pretty easy to achieve high fidelity visuals without having to use genAI as a crutch.
Rik Arends@rikarends
Yea i mean, something the best of the best studio with the best tech, under the worlds bests art direction can achieve doesn't mean everyone can. With AI now many more studios can. Hurrah. Less gatekeeping!
English
@carygolomb Yes I'd much rather have major studios release 1 game every 7 years created by a staff of 1000 people, with constantly cycling layoffs and studio closures, all created on the same 3 or 4 engines.
It's clear the current game creation pipeline is sustainable and good
English
This is not today. But certainly someday you can imagine this eventuality
What's the frame budget on the top part? Gotta crazy fast render. How long did it take to make? Gotta be way faster
To be clear, I really don't want this at all and I hope I'm wrong.
Sergiy Galyonkin@galyonkin
What I really want from DLSS5:
English
@PinaJr @carygolomb I suspect they'll bake it into the game by using more detail than the dummies above and fine tuning
Just enough to inform the conformal generation but not enough to require armies of artists/programmers to render
English
@carygolomb AI people would need to solve the context window issue (i.e. make it massive) so the model could maintain coherence throughout the whole game. Otherwise char appearance and environments would change between scenes.
Genie 3 on TBs of RAM can handle 1min. We'd need ~1000000x more.
English
DLSS 5 got trolled pretty hard but do you notice that out of nowhere almost every post on Twitter now is people defending it.
That's not organic. That's a paid campaign.
English