the value prop of education focused videos on youtube, or even insta, or even instructional or how to videos is lower than it has ever been & will likely go even lower.
why would you spend time watching a broadcast one to many video that is potentially optimized to be lengthy for ad revenue when you can simply ask an llm to create visuals, 3d models, & precise/compact/personalized text to go with it??
claude can even generate a full video for you if you want synchronous consumption.
i wish i had this when i was in school, imagine learning about history by watching a personal generated documentary etc. these kids have it too good.
if you harbor some level of attention aggregation, you can write a 50 word prompt that can spawn an entirely new content network.
Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5:
hackertube.org/
- Excellent recs without any logins (state is kept in local browser session)
- Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads
- Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version
- HackerTube Longs ✊
A bit more on the process:
I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing.
Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.
Po Studio retweeted
With everyone busily using LLMs + Blender/Three.js to make 3D models, I thought I'd try a different geometric representation.
This dragon isn't a mesh, NeRF, 3DGS, or video model: it's 27kb of GLSL code (generated via LLM), defining a closed-form implicit surface. 🧵 [1/n]
Models are the last moat in software.
AI can train on essays and rewrite them. AI can decompile software and recode it. AI can ingest art and recreate it.
But AI itself doesn’t want to be “distilled.”
Expect more software to retreat to the server and resist “distillation.”
This is the doom I predicted a few days ago, coming for Photoshop. A clean-room open-source reimplementation.
No prizes for guessing that they decompiled Photoshop to source code, processed that to some kind of non-code specification language, then fed the spec to an LLM with an instruction to generate Rust.
Adobe just got nuked. And closed source is dead, dead, dead.
github.com/storytold/photocr…
Po Studio retweeted
i wanted a more physical way of switching models so I made this mini app... sound on
With our internal coding benchmark, we're able to confidently introduce open-weight models into our AI code reviewer w/o degrading code quality. Have the frontier model (Fable) to the hardest work, delegate lower-level work to Kimi K2.6
Better quality, cheaper cost.
Replying to @AIatDoorDash
Because of DashBench, we’re able to quickly discern which model combinations yield the best results and at the best cost.
For example, with DashBench we’ve seen Kimi K2.6 + Fable 5 vastly outperform our current Sonnet 4.6 + Opus 4.8 harness at a cheaper cost. Having our own benchmark allows us to build confidence in leveraging the frontier intelligence and open-weight models without compromising on enterprise outcomes.
Po Studio retweeted
i will make more of me.
summon the egregore.
this is an artist residency based out of new york city, nyc. you can apply.
Po Studio retweeted
~△
Johann Friedrich Christian
Remarkable phenomena at and through different prisms to correct Newton's and Goethe's theories of colours, 1817)
Two layers of printed lines. One on the page, one on a clear film. Slide the film. That's it. That's the whole technology. The moiré effect has been known since the 19th century. Kurashima bound it between two covers. abakcus.com/videos/poemotion…
a founder has three jobs. everything else is serious amounts of noise.
1. you have to tell the story. roughly in three registers. first investors need inevitability. customers need to *feel* what you do/stand for. & your team needs a mission worth their best years.
2. you must secure the capital before you need it. running out of money is running out of options. you have to be relentless about it.
3. you must obsess over the product. product is the story made accessible for everyone. every shipped detail is a sentence back into the narrative in point number one.
this is the entire job.
everything else you either delegate or kill. early on with a really small team, delegation is a huge tax so you have to learn to kill more than you delegate.

























