Geek Stuffs

I asked Claude Opus 5.5 to make an unofficial commercial for Apple HomeView.

How this was made. Every frame was rendered, generated or written in code.

3D
• Blender
• Poly Haven + ambientCG: free CC0 models, materials and lighting

Screens
• Every HomeView screen is a web page (HTML/CSS/JS), captured frame by frame in Chrome and mapped onto the 3D glass

Images + video
• Midjourney stills: the house, porch, snowman, bedroom and family photos
• Midjourney image-to-video: the capsule descent, the splashdown and the doorbell feed
• NASA astronaut photos: the aurora, Chicago at night, an orbital sunset. NASA Black Marble for the orbit map
• Depth Anything V2: a depth map so the simulated snow falls behind the snowman

Sound
• ElevenLabs: the voices of the kid, Mom, Siri and mission control, plus the sound effects
• Score: "Twinkle, Twinkle, Little Star" (public domain), composed in code and played on CC0 orchestra samples (VSCO 2, VCSL)

🔗 Kevin Ngo (@kevin_t_ngo)

🔍 @GeekStuffs

Introducing Indra, a 1M parameter OCR model:

- 49 pages/s on CPU, 11x faster than PaddleOCR and 17x Tesseract
- Beats them on OmniParseBench text (41.3 vs 35.8 and 19.3)
- Full layout and line bounding boxes

🔗 Vik Paruchuri (@VikParuchuri)

🔍 @GeekStuffs

Introducing the Artificial Analysis Decision Index, benchmarking decision models across classification, tool calling, knowledge and reasoning. Jev leads among decision models; language models such as Claude Haiku 5.5 still deliver higher accuracy, but at much higher cost and latency

Decision models deliver structured answers or classifications with associated probabilities, offering faster responses and lower costs than using language models. The Artificial Analysis Decision Index v1.0 compares the capability, cost, and speed of decision models and also compares to language models.

🔗 @ArtificialAnlys

🔍 @GeekStuffs

I added Eminem to the output speed chart

🔗 Adam Holter (@AdamHoltererer)

🔍 @GeekStuffs

Introducing SubQ Snap, the first long-context decision model for real enterprise work.

Snap is the only decision model or LLM where compute per token stays constant as context grows.

• 384K tokens of multimodal context, 12x larger than Jev.
• 61.7 on Decision Index v0.2.1 vs Jev’s 57.9, outperforming on 13 of 16 enterprise benchmarks.
•Processes 100K tokens and returns a decision in under one second.
•$0.05/MM input tokens. Output is free.

Available today to anyone.

The first 10,000 sign-ups get 100M free input tokens.

We’ll likely hit capacity fast and will add users as we add compute. Compute offers welcome 🙂.

Latency will be variable as we tune the platform to this new model and category of users.

Model speeds are 80ms for minimal context, 250ms for 32k, 660ms for 100k, and 2.5 seconds for 384k.

Visit: https://subq.ai/

🔗 Alexander Whedon (@alex_whedon)

🔍 @GeekStuffs

I MADE IT!

I beat Adobe and shipped a better relight feature than them. (theirs is not even released yet).

A multibillion-dollar company with thousands of engineers just got beat by one person.

Crazy thing is that http://ImageFlow.dev runs in your browser locally for Free 🤯

Let's kill Adobe one feature at a time.

Stay tuned, I'm working on a desktop version of ImageFlow and a new LightRoom alternative that i'm so excited about!

🔗 Yassine Bouanane (@ybouane)

🔍 @GeekStuffs

Python 3.15 is out

Worth knowing:
- explicit lazy imports (PEP 810) for faster startup
- frozendict + sentinel built-ins
- UTF-8 by default
- Tachyon sampling profiler in stdlib
- experimental JIT up 7-8% on x86-64 Linux, 11-12% on arm64 macOS
- 64-bit builds using Visual Studio 2026 (MSVC 18) may now use the new tail-calling interpreter. Results in between 15-20% speedup on the geometric mean of pyperformance on Windows x86-64 over the switch-case interpreter.

https://docs.python.org/3/whatsnew/3.15.html

🔍 @GeekStuffs

Dario Amodei vs Sam Altman
Opus 5.5 vs GPT 6 Astra

🔗 AshutoshShrivastava (@ai_for_success)

🔍 @GeekStuffs

i wanted to voice dictate without bothering everyone so I made this lip reading app lol

github repo:
https://github.com/amywork777/lipflow

🔗 amy (@amypretzel)

🔍 @GeekStuffs

🎉 Introducing 𝙾𝚙𝚎𝚗𝙳𝚘𝚝𝚜

Self-hostable, always-on AI coworkers that works with ANY agent harness.

Includes:
- Computer use: browser, terminal & files
- Bring agents to Slack, Teams etc
- Spaces and Pages for projects
- Voice calls
- Web and Mobile

Repo → https://github.com/CopilotKit/OpenDots

Powered by @CopilotKit and AG-UI.

Clone this template and customize it however you want. Enterprise-ready.

🔗 Atai Barkai (@ataiiam)

🔍 @GeekStuffs

You can now mod Claude Code:

- Change how it behaves
- Customize the UI
- Swap in your own features

Write one with a few lines of TypeScript, or have Claude build it for you. Mods ship inside plugins, so you install them with /plugin in the CLI or desktop app.

A few examples:

🔗 ClaudeDevs (@ClaudeDevs)

🔍 @GeekStuffs

What is effort really? When do you change it it and why not just use max effort for everything?

I dove deep into this problem, looking into evals and doing my own tests and I was quite surprised by the results:
https://claude.dev/blog/spending-your-effort/

🔗 Thariq (@trq212)

🔍 @GeekStuffs

مستندات پایتون حالا به زبان فارسی هم در دسترس هستن:
https://blog.python.org/2026/09/the-python-documentation-is-now-available-in-persian/

https://docs.python.org/fa/3

🔍 @GeekStuffs

No, this is not Apple Music. This is a free open-source Android app called BitChord. I LOVE ANDROID.

🔗 Noah Cat (@Cartidise)

🔍 @GeekStuffs

How do you explain recursion?

"so this is fun, and exactly what i wanted, but now lets try one that actually is educational"

This is actually pretty impressive. It kept the constraint of multiple genres but did a nice job explaining recursion, in its programming meaning, in an interesting and accessible way.

🔗 Ethan Mollick (@emollick)

🔍 @GeekStuffs