Remember that you are not talking to "the AI", but you are using a powerful text generator to co-create a story about the interaction between a user and the simulacrum of an Artificial Intelligence.
SOFTWARE
-
NVIDIA and Microsoft partner for secure user-controlled AI on Windows
By
–
We are partnering with @Microsoft to enable secure, user-controlled AI on Windows. NVIDIA OpenShell runtime for agents will provide governance tools, policy enforcement, and smart local-to-cloud query routing. Learn more: https://
nvda.ws/4uKWxaB -
GitHub’s Agent Era: AI agents drive evolution beyond code hosting
By
–
🆕GitHub's Agent Era: 14x commits, 200M developers, Copilot’s next act https://t.co/4JxxgKmJOi@github COO @kdaigle explains why AI agents are forcing GitHub to evolve beyond code hosting, how Copilot is moving from autocomplete to CLI, desktop, cloud agents, and ambient… pic.twitter.com/B3afrwBuaG
— Latent.Space (@latentspacepod) 3 juin 2026GitHub's Agent Era: 14x commits, 200M developers, Copilot’s next act https://
latent.space/p/github @github COO @kdaigle explains why AI agents are forcing GitHub to evolve beyond code hosting, how Copilot is moving from autocomplete to CLI, desktop, cloud agents, and ambient -
VLMs learn 3D natively, skipping expert architectures and complex designs
By
–
"VLM^3: VLMs Are Native 3D Learners" This paper shows that VLMs can learn 3D natively. Most 3D vision systems rely on expert architectures, regression heads, heavy augmentations, and task-specific losses. But they show that you can skip the majority of these designs. All they
-
First impressions and code quality with Kombai 2.0
By
–
Tout le monde parle de la qualité du code,
— Jouhatsu | AI Influence Operator (@Jouhatsu_ai) 2 juin 2026
mais les premières impressions restent décisives.
Kombai 2.0 offre peut-être l’expérience de développement IA la plus soignée que j’ai vue depuis longtemps. https://t.co/RYBdcyiEec pic.twitter.com/xqbgt0UWItEveryone talks about code quality, but first impressions remain decisive. Kombai 2.0 perhaps offers the most polished AI development experience I have seen in a long time.
-
Improving DX for OpenAI API, SDKs and platform – seeking feedback
By
–
going to focus the next few weeks on improving DX for OpenAI API, SDKs and platform, what could be better? what are some papercuts and nothing is too small – docs, tutorials, guides, cookbook hit me up with all 🙂
-
Selling the shovels while others burn 100B tokens monthly
By
–
POV: selling the shovels while others blow through 100B tokens a month https://t.co/ozCVNgO1jg pic.twitter.com/JNFR0sUII1
— The Rundown AI (@TheRundownAI) 2 juin 2026POV: selling the shovels while others blow through 100B tokens a month
-
AI identifies cat; bounding box pixel decoding is slow
By
–
An AI can tell you there's a cat in the image. Pointing to the exact pixels is the hard part.
— Satya Mallick (@LearnOpenCV) 2 juin 2026
The reason it's slow: most VLMs spell out a bounding box one coordinate token at a time — some even split "1024" into single digits. But a box's corners are connected. Decode them… pic.twitter.com/eoJu0PiHGUAn AI can tell you there's a cat in the image. Pointing to the exact pixels is the hard part.
The reason it's slow: most VLMs spell out a bounding box one coordinate token at a time — some even split "1024" into single digits. But a box's corners are connected. Decode them -
Benchtalks #2: SWE-bench creator jyangballin discusses ProgramBench and upcoming project
By
–
Benchtalks #2: @jyangballin
, creator of SWE-bench, fresh off ProgramBench (every frontier model: 0% at launch). Out soon with @vincentsunnchen
. -
Codex stores local state database for pulling
By
–
Yes you can… Codex keeps a state database on your local machine that you can pull from