Re-passing activations through the same layers sounds like recurrent computation stitched back in, curious if it still benefits from longer contexts the way dense transformers do.
TECHNOLOGY
-
Chinese AI Models Progress While Remaining Behind Frontier Models
By
–
Excellent article by @RobinRivaton on the OpenClaw fever. While we wonder if Beijing is "catching up" to Silicon Valley, Chinese models are progressing while remaining structurally behind frontier models, but with an industrial diffusion speed that we can't imagine here. Over
-
Top AI Stories: Gemini Mac App, Snap Layoffs, Claude Agents
By
–
Top stories in AI today: – Allbirds ditches sneakers for AI compute
– Gemini lands on Mac with native desktop app
– Audit business with Notion's built-in Claude agents
– Snap cuts 1,000 jobs on AI productivity boosts
– 4 new AI tools, community workflows, and more -
Kea Aerospace Achieves 56,000 Feet Stratospheric Test Flight
By
–
Kea Aerospace Reaches 56,000 Feet in Stratospheric Test Flight
— Ronald van Loon (@Ronald_vanLoon) 16 avril 2026
by @KeaAerospace#Aviation #EmergingTech #Technology #Innovation pic.twitter.com/f65At7Xs62Kea Aerospace Reaches 56,000 Feet in Stratospheric Test Flight
by @KeaAerospace #Aviation #EmergingTech #Technology #Innovation -
Scalable Software Solutions Require Significantly More Code
By
–
The same is of course true of software — to solve a problem in a scalable manner requires a lot more work and a lot more code compared to a simple non-scalable solution.
-
Robotic Mouth Recreates Human Speech Through Physical Mechanics
By
–
Most voice AI tries to fake human speech.
— Pascal Bornet (@pascal_bornet) 16 avril 2026
This robotic mouth tries to rebuild it.
That is what makes this so fascinating to me.
Instead of relying on text-to-speech software or digital voice models, it recreates speech by copying the physical mechanics of how we actually talk:… pic.twitter.com/mtWTbAbTu1Most voice AI tries to fake human speech. This robotic mouth tries to rebuild it. That is what makes this so fascinating to me. Instead of relying on text-to-speech software or digital voice models, it recreates speech by copying the physical mechanics of how we actually talk:
-
Quantum Computing Advantage: Early Adopters Gain Years of Leadership
By
–
Bien dicho. Y los que empiecen a “interpretar” hoy llevarán 3 a 5 años de ventaja a los que esperen. Escribí sobre esto justamente: https://
linkedin.com/pulse/today-wo
rld-quantum-day-heres-uncomfortable-truth-harold-sinnott-q41be
… #WorldQuantumDay #CyberSecurity -
Google adds PC use support to Gemini
By
–
Google is planning to implement Computer Use support for its Gemini desktop app.
I am tuned -
NVIDIA Lyra 2.0 Revolutionizes Large-Scale 3D World Creation
By
–
Holy crap, NVIDIA just made it drastically easier to create large scale explorable 3d worlds.
— Bilawal Sidhu (@bilawalsidhu) 16 avril 2026
No manual stitching of smaller 3d generations like other 3d models. Lyra 2.0 looks pretty damn impressive. pic.twitter.com/GcTGm3oMHeHoly crap, NVIDIA just made it drastically easier to create large scale explorable 3d worlds. No manual stitching of smaller 3d generations like other 3d models. Lyra 2.0 looks pretty damn impressive.
-
Harvest-Now Decrypt-Later Quantum Threat Endangers Encrypted Data Today
By
–
How safe is your encrypted data from a technology that doesn’t fully exist yet? Less safe than you think. The “harvest-now, decrypt-later” threat is real and it’s happening today: https://
linkedin.com/pulse/today-wo
rld-quantum-day-heres-uncomfortable-truth-harold-sinnott-q41be
… #WorldQuantumDay #Cybersecurity