ilaksh

67 ms·

751.
by ilaksh 2y ago
If you don't like what banks do, then learn about the technologies that can replace them today.
752.
by ilaksh 2y ago
So it's almost twice the performance? That's great. But AI could actually easily use 10 times. Anyone heard anything about memristors being in a real large scale memory/compute product?
753.
by ilaksh 2y ago
The next step for computer hardware is mixed-reality glasses like Meta's Orion right? And maybe subtle haptics like rings and wristbands? Or what else is there that would be a progression? I also wonder if we might eventually see easie
754.
by ilaksh 2y ago
Didn't the government recently screw everyone in regards to affordability by declaring that there was no longer a shortage? And therefore compounding was not allowed, only the original drugs? Yet I still see the compounded injections (
755.
by ilaksh 2y ago
https://archive.is/bb7U3
756.
by ilaksh 2y ago
If people normally enter characters in their phone or computer rather than actually handwriting, then there is no reason to keep remembering all of the character details. Just like it doesn't matter that I frequently am uncertain about
757.
by ilaksh 2y ago
I wonder if it's maybe like the new HeyGen video avatars? Or https://tavus.io
758.
by ilaksh 2y ago
It's very hard for me to find fault with this. It seems to be extremely well put together, at least at a high level. But I like to be critical, so I will go outside the stated scope a bit. The national security aspect of AI seems to go
759.
by ilaksh 2y ago
This model seems to be largely uncensored, at least compared to previous releases. For example, it is surprisingly familiar with the physics of interactions between human skin and yogurt. According to a friend who was testing it on replicat
760.
by ilaksh 2y ago
One thing that always seemed important to these discussions is that the serial structure of language is probably not an optimization but just due to the reality that we can only handle uttering or hearing one sound at a time. In my mind the
761.
by ilaksh 2y ago
I think it's not the same but similar.
762.
by ilaksh 2y ago
There are actually multiple open source ML models for 2d to 3d which is clearly what they are doing. The difference with most of them is that this is vectors. There might actually be a similar open source model already. But I think to creat
763.
by ilaksh 2y ago
That article did not help me understand at all. Nothing specific.
764.
by ilaksh 2y ago
I assume the energy savings are significant. Shouldn't this be actually be part of the building code?
765.
by ilaksh 2y ago
Here's my idea https://github.com/runvnc/tersenet
766.
by ilaksh 2y ago
Waymo has been demonstrating fully self driving cars for years. Teslas also do it, and truly necessary interventions are rare these days.
767.
by ilaksh 2y ago
Science came along and made much of philosophy obsolete, especially the parts that are now directly addressed by science and engineering projects. For your comment to make sense it would have to include some context of computer science and
768.
by ilaksh 2y ago
So the C-37B is basically an unmanned Space Shuttle?
769.
by ilaksh 2y ago
I wonder if there is some way to combine this with a language model, or somehow have the language model in the same latent space or something. Is that was vision-language models already do? Somehow all of the language should be grounded in
770.
by ilaksh 2y ago
This shows there are limitations but it doesn't prove they can't be overcome by changing training data. I don't think that LLMs are the end of AGI research at all, but the extreme skepticism of their current utility is mostly
771.
by ilaksh 2y ago
They did that for o1 and o1-preview. Which if you read the paper or do your own testing with that SOTA model you will see that the paper is nonsense. With the best models the problems they point out are mostly marginal like one or two perce
772.
by ilaksh 2y ago
That makes the whole conclusion obviously false. I don't really understand why, but I think we are going to see total denial from a significant percentage of the population all the way up to and past the point where many average mathem
773.
by ilaksh 2y ago
What type of projects are they exactly?
774.
by ilaksh 2y ago
For some reason this web page doesn't render properly in my browser. The main text overlaps the table of contents. I'm using Chrome on KDE (Ubuntu) on a 1920 wide display (minus the side panel). I checked and I don't have the
775.
by ilaksh 2y ago
We have already seen multiple plateaus as we reached barriers in terms of scaling the models the hardware and software. We have continued to innovate and break through them. The latest barrier is scaling training via inference time. New har
776.
by ilaksh 2y ago
It's very obvious that humanoid robots will be quite capable and mass-produced in well under 20 years. Just do some searching on YouTube for humanoid robot and combine that with the progress in AI. The robots are improving quickly. I w
777.
by ilaksh 2y ago
I think decentralized technologies should be part of the discussion when you it comes to replacing technology monopolies. For example, there might be some protocols for cooperative web indexing or search or to provide a common layer that co
778.
by ilaksh 2y ago
The SOTA models can absolutely create system designs very effectively. Have you even tried that? Like literally, let's get someone to randomly propose a few specs. Then we get some developers to try to create a system design. Some have
779.
by ilaksh 2y ago
Have you used the SOTA models? Can you honestly say that any person is easier to work with than an LLM? Have you seen the capabilities of the realtime voice API that OpenAI just released?
780.
by ilaksh 2y ago
I've been programming since approximately age 8, so around 40 years. For the past few years have been focused on generative AI, agents, etc. I have built multiple agent frameworks in Node.js, Rust and now Python. I use AI with my agent

More ›