six

[six @internetphysics]

Incredible scenes tbh

[six @internetphysics]

@ravespecialist nothing new

[six @internetphysics]

tradexyz will basically continue building the breakout app for HIP-n as n → ∞

[six @internetphysics]

breakout product* not app. the asset is the product

[six @internetphysics]

@mert Helion

[six @internetphysics]

had a great idea for a startup launch video, essentially it would be a montage of inspiring clips from films, historical clips, interviews, etc. -- scenes of technological progress that tie your company's identity with themes of techno-optimism and dynamism in america

[six @internetphysics]

@_chenglou maybe @tldraw @notboring

[six @internetphysics]

@gametheorizing That’s a fake parody account

[six @internetphysics]

@yo_itsmatt also that early 2026 growth is probably heavily driven by base ai/bankr szn

[six @internetphysics]

@yo_itsmatt here's fomo (highlighted week is when they added Base)

[six @internetphysics]

I don't get why more people don't just make low budget movies that 10-100x at the box office. You get great financial returns and also cultural significance

[six @internetphysics]

@Tyler_Did_It Prob worth specifying this is a Ventuals specific issue and not Hyperliquid

I guess SPACEX ticker implicitly does that tho

[six @internetphysics]

Specifically wrt media generation. Brain has been able to adapt as quality progresses

[six @internetphysics]

It’s nice that AI was bad before it was good. It would have been too jarring if we immediately went to good AI but we’ve been able to ease into it somewhat

[six @internetphysics]

@kevinakwok yeah ngl ~3x in 25y is way less than i would have guessed from what i see on twitter

[six @internetphysics]

@joinshiftX shift

[six @internetphysics]

MTS got a lot of flack on/post launch but a few weeks in I honestly am seeing the clips a lot more and there are some interesting interviews. Most problems or reasons why something won't work can kinda just be figured out over time

[six @internetphysics]

@tenobrus if they can make enjoyable longform writing with todays models i'd be quite impressed and wouldn't feel annoyed upon finding out

[six @internetphysics]

That's so funny, was just chatting with a friend about how coretext has a quadratic-scaling bug in its complex-script fallback for Tulu-Tigalari (Unicode 16, U+11380+) and four sister scripts added in Unicode 14-16 (Ol Onal, Kawi, Nag Mundari, Toto): the payload (Tulu base + U+034F + Tulu virama + ZWJ) repeated forces one shaping run per cluster, and CGJ between base and virama multiplies runs per cluster from 2 to 3, with timing measured directly via PyObjC on macOS 26.4.1: 2.8 seconds at 25k UTF-16, 4.1s at 30k (already past iOS's 5-second main-thread watchdog), 11s at 50k, and 25.4 seconds at 75k UTF-16 with roughly O(n²) scaling, meaning a 25k payload reliably crashes any iOS 26.5 app rendering it on the main thread, the 75k version hangs the zombie process for 25-40 seconds before cleanup, and the bug cannot be defended by stripping CGJ alone because 14 different variation selectors (U+FE00–U+FE0D) amplify identically

[six @internetphysics]

"Assistants in Visual Studio Code" is doing a ton of work in this chart lol

[Rihard Jarc @RihardJarc]

It's clear that growth for coding tools such as Claude Code has decelerated from the pace it was since the start of the year.

It might be compute- constrain related or due to many clients blowing their full-year AI budgets.

Monitoring this trend very closely with all the alt data. I will provide regular updates.