@valigo@ipsumessepurum One of my favorite things about working in software dev btw is that my peers skew toward being "thinkers" rather than "knowers".
I can't stand "IT guys" because their job skews in the opposite direction. Pure memorization.
@valigo@ipsumessepurum Pseudointellectuals are annoying w/ the "random trivia drop" as an "appeal to authority" tactic.
It's a weak and dishonest shortcut. But midwits don't know any better, b/c they've gone their entire lives simply acquiring others' knowledge and repeating it rather than thinking.
@omarvvvr Rust is an unproductive language for humans writing code by hand, because it's tedious and a pain in the ass.
The only people who used it were basement-dwellers whose time and attention were worthless.
But Rust makes AI write better code b/c it's tedious and a pain in the ass.
@valigo Trying to establish the equivalence of two things when you're unsure of the essence of either is essentially impossible, and anything you say about it is usually purely semantic without any meaningful/consequential substance.
The rare kind of person you can only find online, because they're so loud here, but you never see in person (because they essentially do not exist, like 0.000001% of the population or less).
I refuse to believe these people are real.
None of them are educated, because obviously this IS still GenAI??
And we dont want anybody using GenAI AT ALL. IT DOESNT MATTER WHAT TASKS ITS FOR. IT DOES NOT MATTER IF ITS "MUNDANE" WE DO NOT WANT IT USED AT ALL.
half the journos on twitter: "defund the police actually doesn't mean defund the police"
protest leader: "we don't want no more police. It's that clear."
@supahvee1234 As usual, the more time you spend involving yourself in niche stuff, the more weirdos you will encounter. Weirdos are drawn to odd niche stuff like a fly in the dark is drawn to a light. This is the #1 crusade for weirdos right now. It comes with the territory, unfortunately.
UPDATE: 90/139TPS on single stream DeepSeek V4.1 Flash on 4x DGX Spark, open source!
C1 - 90 tok/s chat 139 tok/s code
368 tok/s at 16 users
1M context
10k tok/s prefill
image support is live too! lets gooo
https://t.co/GGxOcMgOZP
@0xSero@mobilekjk It's definitely real. You might skew towards a little bit of a communist "post money psychosis" if you keep receiving too many handouts though.
Pretending money isn't real is like pretending NFTs were ever worth anything.
Guys I canโt believe it.
I got another 10,000$ donation!
We are at 25,000$ cash and 50k+ in compute credits. I canโt believe this generosity makes my heart melt.
This whole journey has made me so grateful and was humbling
@ay_ushr FWIW I'm building an interface for myself discards the "chat-based interface" and instead builds on how you'd work with whiteboards and mind maps. The AI agent you're working with is preemptive and asynchronous, rather than reactive and synchronous.
@ay_ushr Is it "agent work is..." (e.g. is it inherent to agents) or is it "the current way I'm working with agents is..." (e.g. is it perhaps a consequence of you and/or others not yet discovering the ideal way to work with them yet)?
@liuliu Once that's in a satisfying place, then I'll turn my attention to Inkling... unless Qwen 4 becomes publicly available around the same time and consumes my attention.
My 4x Sparks are still tied up trying to meet/exceed the SOTA inference stack's speeds for DS4.1F, so I can't actually run it right now.
I'm actually pretty interested in the idea of switching between it and DS4.1F, since an expanded panel of benchmarks showed a really interesting pattern where most of DS4.1F's weaknesses are Inkling 2 Small's strengths (e.g., Design Arena), and vice-versa (DS4.1F seems to be a stronger everyday agentic coding model).