Py_nk (@pyrotemis1) 's Twitter Profile
Py_nk

@pyrotemis1

py
pfp : @taxfaberge

ID: 1029112988433768448

calendar_today13-08-2018 21:10:31

836 Tweet

60 Takipçi

179 Takip Edilen

Yuandong Tian (@tydsh) 's Twitter Profile Photo

📢We show that continuous latent reasoning has a theoretical advantage over discrete token reasoning (arxiv.org/abs/2505.12514): For a graph with n vertices and graph diameter D, a two-layer transformer with D steps of continuous CoTs can solve the directed graph reachability

David Baszucki (@davidbaszucki) 's Twitter Profile Photo

Our AI moderation system is built on large, transformer-based models. We distill and quantize these models to keep the system fast and efficient (the models handling our text filters are handling >750,000 RPS). corp.roblox.com/newsroom/2025/…

90ner. (@90nerisgone) 's Twitter Profile Photo

i found an evil wordle clone ⬛⬛⬛🟩⬛🟩⬛🟩⬛ ⬛🟩🟩🟩⬛ 🟢🟩⬛🟩⬛🟩⬛⬛🟨 🟢🟩🟩🟩🟩⬛ 🟢🟩🟩⬛🟩🟩⬛ 🟢🟩⬛🟩🟩⬛🟩⬛ 🟢🟩⬛⬛⬛🟩⬛⬛🟩🟢 🟢🟩🟢 🟢🟩⬛⬛🟩🟩🟩🟩🟩🟢 ✅

ludwig (@ludwigabap) 's Twitter Profile Photo

Galois has to have one of the tuffest mathematician lore of all time dawg >fail math exam at 16 bc examiners cant follow your proofs >submit insane papers to academy but they lose the manuscript twice >get rejected by basically every institution >join the artillery division of a

Galois has to have one of the tuffest mathematician lore of all time dawg

>fail math exam at 16 bc examiners cant follow your proofs
>submit insane papers to academy but they lose the manuscript twice
>get rejected by basically every institution
>join the artillery division of a
Zephyr (@zephyr_z9) 's Twitter Profile Photo

> LLaMA dev at Meta, 2025 > just finished "training" LLaMA 4 Maverick on 120,000 H100s > cost more than Zuck's entire Hawaiian bunker complex > we're gonna own these Chinese open-source posers > Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞) on Twitter posts a link to a paper from "DeepSeek" > some guys from

> LLaMA dev at Meta, 2025
> just finished "training" LLaMA 4 Maverick on 120,000 H100s
> cost more than Zuck's entire Hawaiian bunker complex
> we're gonna own these Chinese open-source posers
&gt; <a href="/teortaxesTex/">Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)</a> on Twitter posts a link to a paper from "DeepSeek"
&gt; some guys from
vas (@vasumanmoza) 's Twitter Profile Photo

If a simple coffee shop has a bot farm with 100s of phones to amplify their message, please consider what a foreign agency or adversarial operator is running on your favorite social media platform. Especially today, please consider that the opinions you read, the calls to

flaw (@flawedaxioms) 's Twitter Profile Photo

go to GRAD SCHOOL. You will SURELY NOT REGRET getting a PHD. it will make you feel COOL and MYSTERIOUS when you tell the guy at the gym that you do ALGEBRAIC GEOMETRY. It will only cost you NINE HUNDRED THOUSAND DOLLARS in missed earnings.

kyle turman (@kyleturman) 's Twitter Profile Photo

Instead of messing too much with c++ styling. I’m running a node server on my mac mini with dashboard html and css, puppeteer takes screenshot and converts to 1bit color png, then e-ink arduino fetches the image from local server, decodes and renders it lol

Justin Skycak (@justinskycak) 's Twitter Profile Photo

Last year I had a conversation with someone who majored in physics at UChicago. He initially started in math & thought he was prepared having taken AP Calculus BC, but he got smacked in the face by the level of abstraction and proof-writing ability that was assumed. He couldn't

Justin Skycak (@justinskycak) 's Twitter Profile Photo

Many people are misunderstanding the message here. Stated more explicitly: Let's be honest with students about where they stand relative to the caliber of skill they hope to join, not just the caliber of skill in their current peer group. Let's not tell students they're well

tomie (@tomieinlove) 's Twitter Profile Photo

I refuse to believe you can just select an element from any arbitrary set. That’s hubris. That’s Biblical levels of greed.

Deedy (@deedydas) 's Twitter Profile Photo

New research shows which AI detectors actually detect work. The #1 has a < 0.5% false positive and negative rate. It works on text passed through "stealth" humanizers too and on new models like (GPT-5, Grok and Sonnet 4.5) too. It's called Pangram.

New research shows which AI detectors actually detect work.

The #1 has a &lt; 0.5% false positive and negative rate.

It works on text passed through "stealth" humanizers too and on new models like (GPT-5, Grok and Sonnet 4.5) too. It's called Pangram.
Dmitrii Kovanikov (@chshersh) 's Twitter Profile Photo

I went on a job interview for a Senior SWE role. They asked me about my experience with Kafka. I told them how in “Metamorphosis” someone can lose their worth in others’ eyes the moment they stop being useful. It’s a chilling reminder of how fragile our sense of belonging can

Taylor Otwell (@taylorotwell) 's Twitter Profile Photo

A few things I've noticed as all devs write code with AI. When you write foundational / architectural code of a new project by hand, you "feel" the code pushing back if your abstraction isn't right. You feel when something is harder than it should be. The code is telling you