MIT CSAIL

@MIT_CSAIL • 345,420 subscribers

MIT's Computer Science & Artificial Intelligence Laboratory (CSAIL). Media Inquiries: [email protected] Check out the latest CSAIL content ⬇️

Shorts

507,583 views

55,108 views

16,626 views

Videos

sweetdream.ai

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Private Show

Join now for exclusive access

Free preview available • Premium content

78,890 views • 1 month ago

232,846 views • 5 months ago

92,997 views • 3 months ago

46,676 views • 1 month ago

145,000 views • 7 months ago

20,953 views • 28 days ago

150,535 views • 9 months ago

61,025 views • 3 months ago

187,423 views • 1 year ago

182,390 views • 1 year ago

36,385 views • 2 months ago

24,899 views • 1 month ago

35,995 views • 2 months ago

57,794 views • 5 months ago

23,815 views • 2 months ago

28,678 views • 3 months ago

38,598 views • 4 months ago

20,180 views • 2 months ago

95,559 views • 1 year ago

82,861 views • 1 year ago

Live Cam

MIT CSAIL

Shorts

MIT researchers taught a robot dog to perceive a 3D world using Neural Volumetric Memory (NVM). This technique lets the bot climb stairs, step over gaps &amp; run autonomously w/a single neural network. Paper: Video: w/UC San Diego

Barrier-grid animations are a clever way to create a moving picture that doesn’t actually move — w/o electronics. MIT's "FabObscura" software helps you produce bolder designs for these displays, which can be attached to objects in homes, stores, &amp; more:

Videos

Watch Anya Live

MIT PhD student Alex Zhang (alex zhang) explains how AI models are "mismanaged geniuses" that could take on a much wider range of tasks. Full video:

Andrej Karpathy, Eureka Labs founder &amp; former Director of AI at Tesla, breaks down how LLMs like ChatGPT "download the Internet." Watch the full beginner-friendly breakdown here:

Free MIT course that covers tools &amp; technologies that are assumed, but rarely taught, in most CS programs &amp; software engineering jobs, like Bash, git, &amp; AI tools:

#otd in 1980 a Japanese computer scientist published a paper proposing the “Neocognitron,” the neural net that directly inspired CNNs: Kunihiko Fukushima’s paper explained back in 1986:

14 sorting algorithms in just 60 seconds. v/tetsuo.agenc //: 👾

The story of Mirai, MIT's AI model that can detect breast cancer years before humans do:

A free MIT course breaking down fundamental math concepts in computer science: Here, MIT prof. Erik Demaine breaks down state machines (Lecture 4). v/MIT OpenCourseWare

A free MIT course about building scalable &amp; high-performance software systems: (v/MIT OpenCourseWare) Here, MIT prof. &amp; CSAIL principal investigator Charles Leiserson discusses why performance matters in computing (Lecture 1).

60 years ago this month, the Fast Fourier Transform was created, a powerful tool for image compression &amp; data analysis. Watch a classic MIT breakdown of FFT, perhaps the most-taught algorithm at the Institute: v/MIT OpenCourseWare

An MIT introduction to computer science &amp; Python programming:

The MIT paper that showed space can be more powerful than time in computers. Full video:

A decade of imaging. Compressed into three months. Here’s how MIT CSAIL &amp; Harvard taught an electron microscope to see like you do:

Is there a "secret sauce" driving the development of AI models? FutureTech researchers at MIT CSAIL found that at the frontier, sheer scale drives LLM performance, but proprietary techniques &amp; algorithmic advances matter more away from it:

A free MIT course breaking down mathematics for computer science &amp; engineering: (v/MIT OpenCourseWare) Here, MIT prof. Tom Leighton, who is also the CEO &amp; co-founder of Akamai, discusses what a proof is (Lesson 1).

An introductory course about deep learning, courtesy of MIT:

This is MIT CSAIL:

The early days of Amazon — a company now valued at ~$2.4 trillion. v/Jon Erlichman

Where can linearized dynamics be trusted w/contact? MIT's "Contact Trust Region" answers this using manipulation theory, differentiable simulation &amp; control. It enables manipulation w/less compute than RL, powering robots to be more energy-efficient:

A breakdown of the simplest neural network from the late MIT Prof. Patrick Winston during his 6.034 (Artificial Intelligence) course. Clip v/tetsuo.agenc //: 👾

MIT researchers taught a robot dog to perceive a 3D world using Neural Volumetric Memory (NVM). This technique lets the bot climb stairs, step over gaps & run autonomously w/a single neural network. Paper: Video: w/UC San Diego

Barrier-grid animations are a clever way to create a moving picture that doesn’t actually move — w/o electronics. MIT's "FabObscura" software helps you produce bolder designs for these displays, which can be attached to objects in homes, stores, & more:

Andrej Karpathy, Eureka Labs founder & former Director of AI at Tesla, breaks down how LLMs like ChatGPT "download the Internet." Watch the full beginner-friendly breakdown here:

Free MIT course that covers tools & technologies that are assumed, but rarely taught, in most CS programs & software engineering jobs, like Bash, git, & AI tools:

A free MIT course about building scalable & high-performance software systems: (v/MIT OpenCourseWare) Here, MIT prof. & CSAIL principal investigator Charles Leiserson discusses why performance matters in computing (Lecture 1).

60 years ago this month, the Fast Fourier Transform was created, a powerful tool for image compression & data analysis. Watch a classic MIT breakdown of FFT, perhaps the most-taught algorithm at the Institute: v/MIT OpenCourseWare

An MIT introduction to computer science & Python programming:

A decade of imaging. Compressed into three months. Here’s how MIT CSAIL & Harvard taught an electron microscope to see like you do:

Is there a "secret sauce" driving the development of AI models? FutureTech researchers at MIT CSAIL found that at the frontier, sheer scale drives LLM performance, but proprietary techniques & algorithmic advances matter more away from it:

A free MIT course breaking down mathematics for computer science & engineering: (v/MIT OpenCourseWare) Here, MIT prof. Tom Leighton, who is also the CEO & co-founder of Akamai, discusses what a proof is (Lesson 1).

Where can linearized dynamics be trusted w/contact? MIT's "Contact Trust Region" answers this using manipulation theory, differentiable simulation & control. It enables manipulation w/less compute than RL, powering robots to be more energy-efficient: