Dev meme (@devs_memes) header
Dev meme (@devs_memes) profile picture

This account exists because the function is not a function IT! MEMES! American Welding Society!

Planet Earth
Joined February 2019
Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)Photo by Dev meme (@devs_memes)
WTF Ok, forget about memes for a second NIST is deprecating ECDSA/EdDSA/RSA as soon as 2030 and disallowing its usage since 2035 nvlpubs.nist.gov/nistpubs/ir…
2
2
1
35
5,460
"They didn't agreed to pace frontier"
3
83
Dev meme retweeted
Crazy prediction: in coming few years GPU-rich companies like #SPCX, #GOOG, OpenAI, Anthropic, DeepSeek, Moonshot, Zhipu AI will start introducing token allowance based on amount of shares in possession, likely verified on chain
1
111
Dev meme retweeted
🤣
1
2
16
293
Lets gooo!
Every AI hedges: "ChatGPT can make mistakes." I made a Chrome extension that drops the "can." 🤡 ChatGPT, Claude, Gemini, Grok — the footer finally admits it. Plus a button that tells the model to double-check itself. 🔗 links in replies
1
115
When we joke that Linus was prompting programmers 20+ years ago like they all LLMs nobody is joking actually
1
2
97
PMs? you mean backward deployed engineers?
2
89
Sandboxes are just escape rooms for llms
3
98
If this is not most cursed cyberpunk which we live in idk what it is then
1
1
111
Forget loops, we vibe solving frontier math problems
88
Full new update of devmeme! Go try to find your favourite memes from previous years <3
1
214
Finally! Proper benchmark that I can fully agree with That’s what makes Fable special: ratio of bullshit output to result, with cost being actually low, when evaluated per-task . And less tokens for your brain to consume, less burnout, more to achieve
I benchmarked Fable 5 Low effort vs GLM 5.2 High effort and Sonnet 5 High effort using @VulcanBench 🖖 Evals were on 10 real engineering tasks from flask, aiohttp, sqlglot + more. All post-training-cutoff, so nothing memorized. All three scored 80%. The costs were $2.27 vs $8.41 vs $15.81. And the wildest thing is which model was the most expensive, and it wasn't Fable 👀 Results in first comment below 🔽
101