"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
TECHNOLOGY
-
Debating benchmarks all day, users care about experience
By
–
we can debate benchmarks all day users care about experience
-
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 🙂
-
Factory’s auto model selection outpaces OpenAI and Anthropic
By
–
Factory now auto selects the right model for you…I don’t see how Anthropic or OpenAI can compete with this.
— Nathan Lands (@NathanLands) 2 juin 2026
Models will continue to be good at different things, with different cost profiles. And that will change more rapidly over time.
Now you don’t have to worry about that. https://t.co/ChWaOWnzv3Factory now auto selects the right model for you…I don’t see how Anthropic or OpenAI can compete with this. Models will continue to be good at different things, with different cost profiles. And that will change more rapidly over time. Now you don’t have to worry about that.
-
Mustafa Suleyman predicts three OOM jumps in training compute by 2029
By
–
https://t.co/LizAzIh7dN pic.twitter.com/VVreg7R7OW
— Bojan Tunguz (@tunguz) 2 juin 2026At Microsoft Build today, Mustafa Suleyman predicted three more OOM jumps in the amount of training compute between now and summer 2029.
-
Taiwan powers global AI factory buildout with accelerated computing and AI agents
By
–
Taiwan isn’t just building AI infrastructure, it’s powering the world’s AI factory buildout. 👏
— NVIDIA (@nvidia) 2 juin 2026
From fabs to factory floors, leaders like TSMC, Foxconn, QCT, Pegatron, and Wistron are using accelerated computing, digital twins, and AI agents to transform how AI systems are… pic.twitter.com/E50uZG0t3lTaiwan isn’t just building AI infrastructure, it’s powering the world’s AI factory buildout. From fabs to factory floors, leaders like TSMC, Foxconn, QCT, Pegatron, and Wistron are using accelerated computing, digital twins, and AI agents to transform how AI systems are
-
NemoClaw on DGX Spark: AI agent from unboxing in minutes
By
–
From unboxing to AI agent in minutes. Getting an agent running used to mean sourcing a model, configuring an inference backend, installing a runtime, and wiring it all together. The new NemoClaw install path on DGX Spark replaces that with a single command. DGX Spark also
-
AI crunches SDR and real estate manuals to understand human psychology
By
–
Human psychology is a complex thing. It can be actively reasoned about. We now have surprisingly powerful reasoning machines. They have crunched through many sources of data, including SaaS SDR training manuals and real estate agent how-to books.
-
Quadratic cost kills kernel machines and transformers
By
–
Quadratic cost killed kernel machines, and will kill transformers.
-
Microsoft AI technical write-up praised for rare detail and quality
By
–
Kudos to the team on a really good technical write-up, though, which a lot of AI labs are no longer providing. https://
microsoft.ai/wp-content/upl
oads/2026/06/main_20260602_2.pdf
…