You will only be right for a few more weeks. And I never look at the road anymore. So for me it's already Level 4.
AI
-
New AI Computers Enable Data Scraping of User Likes
By
–
But you can use one of the new AI computers to scrape your likes.
-
Leading AI Researcher Secures Historic Seed Round at AMI Labs
By
–
You're right, I don't have a job. I have TWO jobs:
1. Jacob T Schwartz Professor at the NYU Courant Institute. 2. Executive Chairman at AMI Labs, for which we just raise the biggest seed round in history outside the US. -
Multimodal LLMs Fail at Car Detection Task Compared to YOLO
By
–
I tested Opus 4.7 on a simple car detection task.
— Satya Mallick (@LearnOpenCV) 20 avril 2026
→ Took 5 minutes
→ Missed multiple cars
→ Pointed at blank spaces
→ Bounding boxes were even worse
Codex did better (24 cars, accurate points) but still took 3 minutes.
YOLO does the same thing in 30ms.
Multimodal LLMs are… pic.twitter.com/DiHGsYpOG2I tested Opus 4.7 on a simple car detection task.
→ Took 5 minutes
→ Missed multiple cars
→ Pointed at blank spaces
→ Bounding boxes were even worse
Codex did better (24 cars, accurate points) but still took 3 minutes.
YOLO does the same thing in 30ms.
Multimodal LLMs are -
Low Rank Dense Bubbles in Machine Learning
By
–
yea i think it's more about "low rank" here being close to the centroid of a local dense bubble
-
1D Ordered Tokens Improve AR Image Model Test-Time Search
By
–
“(1D) Ordered Tokens Enable Efficient Test-Time Search” Most AR image models generate 2D grid tokens patch-by-patch, so partial generations are hard to verify. This paper shows that 1D coarse-to-fine tokens make search much easier, with early tokens encode global semantics,
-
Autonomous Coding with Cline: DevTalks Webinar
By
–
Time for another DevTalks This one: autonomous coding with @cline
. Faster inference, less friction, full control. April 30 | 11am PT | 2pm ET https://
sambanova.ai/webinar-ai-ass
isted-coding-workflows-with-cline.bot-sambanova?utm_source=x&utm_medium=organic&utm_campaign=developer&utm_content=events
… -
From AI Experimentation to Military Mission Impact
By
–
We had a great time with Carahsoft at the Aberdeen AI/ML Tech Showcase. We enjoyed conversations with the Army and DoW community about what it really takes to move from experimentation to mission impact. #MissionDrivenAI #partnerships #Army
-
Google loses coding agent race lead to competitor
By
–
Google was in pole position for the coding agent race, but as usual they let someone else run away with it.
-
Kimi 2.6 Code: Open-Source Claude Code-Like Terminal Agent
By
–
Introducing Kimi 2.6 Code.
— Pietro Schirano (@skirano) 20 avril 2026
A Claude Code-like terminal experience built specifically for Kimi K2.6, effectively making it one of the most powerful open-source coding agents on the planet.
Simply bring your API key and use /login.
Repo here 👇 pic.twitter.com/x3nYuiU3lXIntroducing Kimi 2.6 Code. A Claude Code-like terminal experience built specifically for Kimi K2.6, effectively making it one of the most powerful open-source coding agents on the planet. Simply bring your API key and use /login. Repo here