After Google acquired my AR startup, I started thinking about the next interface, this time for AI. Text and voice work fine, just like DOS did before the Mac. But nothing’s more intuitive than face to face.
After two years of research, today we’re launching Atmee v2: a real-time avatar model that makes live video agents feel truly life-like.
• Any character from 1 image + 10s of audio
• Contextual expressions, gestures, even interacting with the environment
• Live video calls, voice only, or video generation
At its core, Atmee v2 is a video foundation model, it can render anything, but it’s built to excel at characters, which is what video models like Seedance 2.5 and Wan 3.0 are mostly used for. In comparison:
Atmee v2
• Up to 50x faster
• 10-100x cheaper
Build tutors, sales coaches, support agents, companions, game characters, even whole living worlds. The interface after text and voice starts today.
Build your own harness, folks.
This is absolute banger paper from NVIDIA on self-evolving agent harnesses.
(bookmark it)
They introduce SoL-Pi which cuts token traffic by nearly half.
And it matches its baseline harness on GPT-5.6 Sol and Opus 5.
More details below:
Instead of tuning a harness by hand, they run auto-research loops at the harness layer across many repository-derived and verifier-driven environments, keeping only the mechanisms that survive selection.
Four mechanisms survived:
> Action Fusion changes how actions execute
> Online Context Compact handles compaction during a run
> ObservationPack reshapes observation handling
> Evidence-Preserving Reducer covers delegated reading
On the 51-task EdgeBench evaluation, the savings translate to about a third off API cost. In dollars that is an estimated $8.75 to $13.50 per hour against native Codex and Claude Code harnesses, and $4.36 to $5.71 against the baseline harness.
Because the search runs across many environments rather than one, the retained mechanisms keep working outside the setting that produced them. Code is on GitHub under NVlabs.
Paper: https://t.co/1x26LzuE6d
Chat with Paper: https://t.co/kygTc5XLFB
A frame-by-frame stop-motion recreation of the Beastie Boys’ iconic “Sabotage” video by independent animator Corduroy Frames.
Five years of work. 4,340 hand-animated frames made from card stock, foam, cardboard, tin foil and cotton balls. No AI. Just pure obsessive craft.
@politvidchannel If it's true that there was interference...maybe so, maybe no....I have no idea, but if I were gaming it out, maybe it wasn't an effort to get Trump in, maybe the effort was to get a relatively unknown Vance into office, and trump was the vehicle. No Accusation, just a thought.
If you're still using Swashbuckle in .NET 10, here's what you missed.
Swashbuckle hasn't had regular updates in over a year.
So, Microsoft dropped it from .NET 9+ templates and built their own.
The replacement: `Microsoft.AspNetCore.OpenApi` — built-in, lighter, maintained.
For a UI: Scalar > Swagger UI. It's faster, better looking, and supports dark mode out of the box.
Migration takes 5 minutes: remove Swashbuckle, add `https://t.co/SUgtYkwEkM.AddOpenApi()`, done.
Read article to help you migrate -https://t.co/eGz3clMjAu
Repost this to help a fellow developer.
Ok....so having now read the debate in the comments about lines of code and features...so "last year" btw....but I digress...what I'm interested in is the architecture, feature sets, and choreography of these systems, and considering many of us don't really "code" anymore, why don't we start sharing specs and architectures rather than code these days? Just feel like it would be more useful, no?
Stanford researchers just published a prompting technique that makes today’s LLMs behave like better versions of themselves.
It’s called “prompt ensembling” and it runs 5 variations of the same prompt, then merges the outputs.
Here’s how it works 👇
I've been using Claude Code for several months, been pretty happy with it, no real complaints until I tried Factory Droid over the past few days.....it just hits different, I kinda like it.
@BernieSanders Oh yeah....and all that should go straight to Social Security, and retraining programs, not the general fund.....in addition to current levels, obviously.
@BernieSanders Tax LLM tokens. Done
Tax offshore workers for full SS withholding. Done.
Pretty straight forward. Every time a business tries to circumvent American labor....make them pay.