Frosty

[Gemini Notebook @Gemini_Notebook]

☎️ Calling all students:

You can now access hundreds of textbooks directly in Notebook, for no charge. From publishers like S Chand Publishing, OpenStax, and Open Textbook Library— browse, select and upload them as sources to ace all of your upcoming exams (no more excuses, ok?)

[Beamsters @Beamsters1]

Artificial Analysis v4.3.2 to v5, AI models re-ranking step by step.

[Frosty @Frostifey]

@Beamsters1 This is GOLD man.

[Frosty @Frostifey]

@theo @amritwt tibo and poteto and I (theo) might be the best devrels of all time

[Frosty @Frostifey]

@osanseviero Perfect fit for 12GB vram at Q4 and context

[Frosty @Frostifey]

@DogukanUrker @CaryPalmerr @PrismML Hey mate, leave the garbage models and 2 bit quants and models too big to run on 12GB VRAM. Focus on good quants above Q3_K_XL and actually good models we would want to run to do some tasks on our systems. merely running models and tallying that this model runs is not useful.

[Aakash Gupta @aakashgupta]

The typical American guy working full time makes about $35 an hour, so every hour he spends on the couch with his best friend quietly costs him $35. That's the real reason men can't hang out without golf or the game on, and a Swedish economist figured it out back in 1970.

His name was Staffan Linder. He wrote a book called The Harried Leisure Class, and the idea is pretty simple once you hear it.

Every hour you have is worth whatever you could've earned with it. As a country gets richer and wages go up, your free time gets more expensive right along with your work time. Linder said people react by trying to squeeze more out of every free hour, and the easy way to do that is to stack stuff on top of it. So the hour turns into a round of golf, a game on the big TV, a fishing trip, a new grill you want your buddy to see.

That's what Trevor is describing with "the third thing."

When a guy texts "want to go fishing Saturday?" he's attaching something to the hour so it feels worth spending. The fishing gives him a reason to be there. Two guys on a couch just talking feels like nothing got done, so if you grew up somewhere wages are high, your brain files it as wasted time. Even though it's probably the best thing either of them will do all week.

It also explains why Trevor's friends from the township in South Africa and from Trinidad don't have this problem. Trinidadians even have a word for it, liming, which basically means hanging out with zero agenda. An afternoon of doing nothing in Soweto costs a fraction of what it costs in Manhattan, so it never got squeezed out of anyone's week.

Now do the math on an actual close friend. A University of Kansas researcher named Jeffrey Hall found it takes more than 200 hours together before someone becomes a close friend. At $35 an hour, that's almost $7,000 worth of time per friend. Some part of your brain runs a version of that number every time "just come over" feels like it needs a reason attached.

Later in that same interview, Trevor tells a story that's basically Linder's whole book in one afternoon.

Once he was hosting The Daily Show, a theme park offered him and his friends a pass to skip every line. They rode everything a few times over in about three hours, in a park that usually takes all day. He came home with a headache and his friends were dehydrated. They'd barely talked, and it was the most depressing theme park trip they'd ever been on.

That night they tried to figure out what went wrong, and it hit them that the lines were the part they loved. Standing in a queue for an hour meant there was nothing to do but hang out and talk.

That pass did to one afternoon what a big salary does to your whole social life. It got rid of all the dead time, and the dead time was what they came for.

[Frosty @Frostifey]

@Internetgeist @brombinmirko Fsearch is not fast indexer like everything. Also Everything UX, wildcards and operators support is better.

[Frosty @Frostifey]

@naks0dn @brombinmirko I think you haven't used everything on windows.

[NIK @ns123abc]

What's your private key? 1?
>...what? No?
2?
>It's 384 bits you dumbass. That's 35 orders of magnitude larger than the estimated number of atoms in the known universe.
...3?
>[visibly sweating] Th-this is ridiculous...

[Tibo @thsottiaux]

@JustLingonberry @sama is this true

[Jeff_synthesized @JeffSynthesized]

Action movies will never be the same again.
4K YT https://t.co/MX7rwRWliI

[Frosty @Frostifey]

@TuckerGoodrich simply add it before the fold.

[Frosty @Frostifey]

@hqmank A Guy at that position and being how Honest he is, is highly commendable.

[dax @thdxr]

they've been non stop trying to attack our systems for the past 48 hours

a side effect is everything is getting hardened but damn we basically need agents going 24/7

because that's what they have

[Dan Shipper @danshipper]

Codex is one of the most important untold stories in the history of technology.

A year ago, it was a scrappy effort in a quiet corner of @OpenAI. Its cloud product hadn’t found PMF, and the team was starting over from scratch.

Less than 12 months later:

- Codex swallowed the flagship ChatGPT desktop app from the inside (it literally runs as com.openai.codex on your Mac)
- Codex and ChatGPT for Work hit 40M+ weekly users across engineering and knowledge work
- Its leader @thsottiaux was elevated to run all of OpenAI’s core products, including ChatGPT, Codex, and the API

How did they pull it off?
They trained their own dedicated model line, separate from ChatGPT, with researchers and product builders sitting on the same team

They built four different products in a single year (cloud → terminal → IDE sidebar → a desktop app with no code editor at all) as every model leap made a new form factor possible

They used Codex to build Codex—so a handful of people could prove a new product in weeks instead of arguing for it in meetings

I had a front-row seat to the whole arc—from their failed Cloud launch, to OpenAI’s offices the night before GPT-5 dropped, to the day Codex took over ChatGPT.

And I spent months interviewing @gdb, @thsottiaux, @embirico, @mia_glaese, @aambrosino, and @edbayes to get inside how it all went down.

Here is the full, behind-closed-doors story of the team that remade OpenAI:

https://t.co/Ai7lTOJY5r

We’re pushing the boundaries of how stories like this are told.
This piece is completely agent-native—designed to be read alongside your agent inside ChatGPT’s in-app browser:

- Every interview transcript is included in full: Click any quote and you (or your agent) can see the exact conversation it came from.

- Built for your agent to explore: Ask your agent to pull out product lessons for your team, cross-examine the transcripts, or find insights I missed when writing the piece.

It’s a story about the future of software—built to be read with an agent:

https://t.co/Ai7lTOJY5r

[OpenCode @opencode]

DeepSeek V4.1 Flash is now free in OpenCode

[Tibo @thsottiaux]

@poteto @theo I made it alive

[NVIDIA RTX Spark @NVIDIARTXSpark]

llama.cpp now has a Windows app 🦙

Llama Windows makes it easy to run models locally, with optimizations for NVIDIA RTX Spark.

📥 Browse and download models from Hugging Face
💬 Chat with local models right from your system tray
🔒 Keep your prompts on your PC