Viewing profile — nl

> Only for China, of course HP, Acer and Asus are now using them https://asia.nikkei.com/business/china-tech/hp-asus-and-acer...

I think this is a great story. For those who aren't aware, that strip of land is probably some of the most valuable residential land on earth. HK: $200,000+/ sq m (for premium prop…

I sent the blog post to someone right now because it had exactly the kinds of details they care about, with the appropriate level of technical details.

Do you actually think that this group of founders will have trouble finding domain experts wanting to work with them?

This is a very bad way of thinking of it. Small LLMs have clues about real knowledge but only surface level answers will be accurate.

You can tell someone in native on LessWrong by their ability to use incredibly obtuse language to say simple things. It's just in-crowd signalling. > cunningham-esque laws Cunningh…

https://huggingface.co/blog/lerobot-release-v060#molmoact2 Read the command line prompt: --task="pick up the red cube"

https://huggingface.co/blog/lerobot-release-v060#molmoact2 Read the command line prompt: --task="pick up the red cube" There is a gif directly below it. This is a completely open s…

> Michael O’Kelly in the Google+ comments pointed out that the scheme was also in conflict with option-value/learning I miss Google+ (and even more Google Buzz that came before it,…

Because they need internet access to eg search for things.

The actual project README appears more human-written and is actually pretty interesting: https://github.com/HUANGCHIHHUNGLeo/claude-real-video

That spinning person one would make an interesting benchmark. The model clearly has a strong prior that human heads go upright.

> Ironically, it died around the time LLMs/AI started becoming good. What? I think maybe you are confusing Itanium with something else? Development on itanium stopped in 2013: > On…

> In the interest of perspective, can anyone (perhaps playing devil's advocate) give one? I have numerous cases where Sol failed and only Fable could solve a problem. For example y…

Wow the Opus version is a lot more functional (try clicking some links). I'm quite surprised at the difference.

That's not how you measure inference costs though. That includes training.

> Action LLMs work by generating text underneath This isn't true. Obviously there is a lot of variety in architecture, but in the prototypical example there are vision and language…

Why? Have you ever tried reproducing even a small neural network exactly if you train on GPUs on more than one machine? I have and it is pretty close to impossible, and I'd argue a…

[dead]

You fine-tune the open-weight model without access to the original weights. If you want to train it from scratch you need data, yes. But presumably if you are doing that there is a…

> Revenue: $3.7 billion > Cost of Revenue: $2.65 billion That's how standard accounting rules for public companies would measure it.

Margins means exactly what the poster you are replying to implies: that in the US Tesla has high margins (the difference between cost of good sold and what you get for them) becaus…

>But how does LLMs help in making chat bots better, help with this "multi-sensory" data. All three offerings from the linked blog post are either Vision LLMs or Vision/Action LLMs.…

Using LIDAR is of course the perfect example of the bitter lesson. More data makes for better outcomes.

This seems like a post by someone who hasn't really ingested the bitter lesson. Eg, even LeRobot (without proper fingers) can fold clothes now: https://www.youtube.com/watch?v=dPe9…