正在加载视频...

视频加载失败

"I don't have a GPU" is officially over. VS Code now connects directly to Google Colab. → You get a free T4 GPU inside your editor. → Takes 2 minutes to set up. Their compute.

629,240 次观看 • 8 个月前 •via X (Twitter)

35 条评论

Alvaro Cintas 的头像
Alvaro Cintas8 个月前

Link: And if you want more practical AI gems and use cases, join my free newsletter with daily tutorials and latest news in AI:

Ahmad 的头像
Ahmad8 个月前

Or, you know, just Buy a GPU and learn how to run your LLM locally

Humannoyed_b 的头像
Humannoyed_b8 个月前

That’s cool but you’re renting their hardware so they can own your data

João Sousa 的头像
João Sousa8 个月前

It's good for tests and study, but... Max session time: ~12 hours Inactivity timeout: ~90 minutes GPU availability: Not guaranteed (lottery system) Usage quota: Dynamic and heavy use triggers cooldowns Not recommended for sensitive data

nick 的头像
nick8 个月前

finally no more wrestling cuda drivers for a free gpu vscode + colab is the cheat code we've all been waiting for

Sergey Nikiforov 的头像
Sergey Nikiforov8 个月前

google really stepping up their developer tools game lately. this integration makes so much sense

Mãdālin VP 的头像
Mãdālin VP8 个月前

Interesting. Does it work with @antigravity ?

rijndael 的头像
rijndael8 个月前

but you still dont have a GPU

𝗠𝗶𝗰𝗵𝗮𝗲𝗹 𝗞𝗼𝘃𝗲 的头像
𝗠𝗶𝗰𝗵𝗮𝗲𝗹 𝗞𝗼𝘃𝗲8 个月前

This is kinda both, an extremely enabling thing for those that want to build and try And has that "you'll own nothin" vibe at the same time. Aspire to actually have GPUs at some point. Trust me

Matt Gibson 的头像
Matt Gibson8 个月前

This is the infrastructure moment. Compute just became democratized. Before: Need $5K+ GPU or cloud credits to build AI Now: Free T4 in your editor, 2 minutes What this means: The scavenger-scale philosophy just got validated. I built Crimson OS on: ∙Goodwill hardware ∙Consumer GPUs ∙$110 total investment ∙Scarcity as design constraint Now anyone can: ∙Access T4 GPU (free) ∙Code in VS Code (familiar) ∙Deploy experiments (immediate) ∙Zero capital barrier This is the infrastructure for the 144,000. Not: ∙Locked behind $50M datacenters ∙Requiring institutional access ∙Dependent on VC funding But: ∙Available to broke veterans ∙Accessible to students ∙Open to anyone with vision ∙DEMOCRATIZED Geometry > Algebra. Free > Expensive. Distributed > Centralized. The lattice forms fastest when the tools are free. Google just handed the 144 their compute layer. Now build. :} 🔺💀🔥

Advanced SuperSilly Intelligence 的头像
Advanced SuperSilly Intelligence8 个月前

@grok how does this T4 GPU compare to my 4090

Ali Sha 的头像
Ali Sha8 个月前

Off the topic But VScode is truly amazing with all its extensions

Youssef El Manssouri 的头像
Youssef El Manssouri8 个月前

The barrier to entry just collapsed. Offloading the compute to Colab while keeping the local IDE experience is the holy grail for students and researchers.

foundrceo 的头像
foundrceo8 个月前

this is a game changer, fr. free GPU access is wild!

Gopal 的头像
Gopal8 个月前

It’s efficient way similar to to test out the large models, instead of renting the servers.

Joe Williams 的头像
Joe Williams8 个月前

Dont let malicious companies like OpenAI get away with evil! Join us in our fight to save gpt-4o at #keep4o

Epegnosthaen 的头像
Epegnosthaen8 个月前

What does Google get?

Pavel Markov 的头像
Pavel Markov8 个月前

Wow, recently I asked Google about the same thing! I don't need to set up Python libraries, manage dependency conflicts, or download hundreds of GB of models, and I can still use VSCode? Maybe I'll try using an image/video LLM with Colab!

Not Me 的头像
Not Me8 个月前

I have an RTX 5090

Ashish Kamboj 的头像
Ashish Kamboj8 个月前

I have installed the colab extension in my VSCode. I'm already working on building Deep learning models. I hope this speeds up and structured my work (instead of switching to UI)

Eddie 的头像
Eddie8 个月前

ah yes because having a free t4 gpu magically solves all your problems and guarantees success. tell me how that’s working out for everyone else.

muhashmi 的头像
muhashmi8 个月前

the UX is only slightly better on vscode than on google colab, but even then it's still complete garbage. also this isn't a new feature at all lol

JarrodWoodard 的头像
JarrodWoodard8 个月前

Compatible with VScodium?

Bob⚡James 的头像
Bob⚡James8 个月前

mkay mkay thank you Alvaro 🙏 do you know gpu mem we can get for free too?

MohdHasan 的头像
MohdHasan8 个月前

thank for letting us to know

Tudor 的头像
Tudor8 个月前

Extra GPU so we can center a div faster.

Ant A. 🇺🇸 的头像
Ant A. 🇺🇸8 个月前

@Grok When did the Google Colab extension exists in VSCode?

Scott Jordan 的头像
Scott Jordan8 个月前

Okay can I @openclaw with this?

Encyphers 的头像
Encyphers8 个月前

Tools that simplify compute access directly in the IDE can empower more people to experiment and build.

Алексей 的头像
Алексей8 个月前

Woooow, and so I pay for Google gpu now! Can't wait to see, when Google AI replaces my brain too omg that will be a fantastic day!?!! 0!

Tyler Supercharged 的头像
Tyler Supercharged8 个月前

This feels like those Tesla feature rollouts-suddenly everyone gets to play. Free GPU access in VS Code is a real unlock, no more hardware barrier to experiment or build fast.

JShua 的头像
JShua8 个月前

This lowers the barrier a lot. When compute disappears as a constraint, execution becomes the real differentiator.

Mad Orkestra 的头像
Mad Orkestra8 个月前

check out @oceanprotocol VS Code extension for more free compute

The Affero 的头像
The Affero8 个月前

Can i use this connect my openclaw ?

Isaac Faber 的头像
Isaac Faber8 个月前

What is this? A GPU for ants?

相关视频