ilaksh

30 ms·

511.
by ilaksh 1y ago
As someone who works on his own open source agent framework/UI ( https://github.com/runvnc/mindroot ), it's kind of interesting how announcements from vendors tend to mirror features that I am working on. For e
512.
by ilaksh 1y ago
The web pages are working properly on my 5-6 year old laptop running Ubuntu and Chrome. Maybe it doesn't work in Safari or on Macs?
513.
by ilaksh 1y ago
Hm. There are some things that banks do then that math can't do. But advanced cryptocurrencies have smart contracts that allow DeFi systems to handle many typical banking functions. The primary things that people use them for, to store
514.
by ilaksh 1y ago
Maybe try to make a list of 1 or 2 things instead of a mile.
515.
by ilaksh 1y ago
I don't care how many times I am violently buried on this site for mentioning the word -- but cryptocurrency makes traditional banking obsolete. Or should have.
516.
by ilaksh 1y ago
some people cheat by using web components but with Lit Elements. https://lit.dev/ . I use it with raw JS without any bundling.
517.
by ilaksh 1y ago
If they are an average person who wants to talk something out and get practical advise about issues, it is generally not safety critical, and LLMs can help them. If they are mentally ill, LLMs cannot help them.
518.
by ilaksh 1y ago
SOTA models can actually handle complexity. Most of the discussions I have had with my therapy agent do have a lot of layers. What they can't handle is someone who is mentally ill and may need medication or direct supervision. But they
519.
by ilaksh 1y ago
I've built a therapy agent running my own agent framework with Claude 3.7 based on research into CBT (research aided by my agent). I have verified that the core definition and operation of therapy sessions matches descriptions of CBT t
520.
by ilaksh 1y ago
The leading edge models are trainable via instructions. That's why agents are possible. Many online therapy or therapy companies are training or instructing their agents in this domain.
521.
by ilaksh 1y ago
The biggest thing that has improved is the intelligence of the models. The leading models are much more intelligent and robust. Still brittle in some ways, but totally capable of giving CBT advise. The same way end users ever get to use a t
522.
by ilaksh 1y ago
I am not talking about a layperson building their own therapist agent from scratch. I'm talking about an expert AI engineer and therapist working together and taking their time to create them. Claude 3.7 will not act in a default way g
523.
by ilaksh 1y ago
Which model exactly? What type of therapy/prompt? Was it a completely dated model, like in the article where they talk about a model from two years ago? We have had massive progress in two years.
524.
by ilaksh 1y ago
The context is that WWIII is basically a foregone conclusion? If there is a hot war between the US and China, that proves that humans aren't fit to control the planet. The only way this makes sense for people is if they are racists dee
525.
by ilaksh 1y ago
How does this compare to tinc?
526.
by ilaksh 2y ago
Technically it should be the company that is responsible.
527.
by ilaksh 2y ago
I use AI as much as possible for programming, but the specific wording "replacing" is not quite there yet. The confusing thing for people is it probably will be at full replacement level within a couple of years. The leading edge
528.
by ilaksh 2y ago
Different groups or individuals are going to have different reactions. Most of the reactions to AI art are similar to the reactions for art and actually most things in general. They are based on popularity or fashion more than any real judg
529.
by ilaksh 2y ago
Wouldn't a human normally make the D&D characters though? Also, do they hate all AI art, or just art where they can tell because it looks like AI art?
530.
by ilaksh 2y ago
It makes me think that maybe our visual perception is similar to what this program is doing in some ways. I wonder if there are any computer vision projects that take a similar world emulation approach? Imagine you collected the depth data
531.
by ilaksh 2y ago
This seems incredibly powerful. Imagine a similar technique but with productivity software. And a pre-trained network that adapts quickly.
532.
by ilaksh 2y ago
People just don't know the difference between popularity and merit. Similarly, they don't know the difference between someone who is successful or good at what they do versus one who makes a lot of money.
533.
by ilaksh 2y ago
Anyone else explored a plugin approach to bundle tools with the client? I started down that path long before MCP was a thing. I do plan to add MCP and A2A support at some point. But for my program it's generally not necessary. You just
534.
by ilaksh 2y ago
Amazing. If you have more RAM or something, can it go faster? Can you get even more speed on an H100 or H200?
535.
by ilaksh 2y ago
It's using the new gpt-4o, a version that's not in the API
536.
by ilaksh 2y ago
in the api or on the website?
537.
by ilaksh 2y ago
wait, o4-mini outputs images? What I thought I saw was the ability to do a tool call to zoom in on an image. Are you sure that's not 4o?
538.
by ilaksh 2y ago
Did you try the same exact test with o3 instead? The mini models are meant for speed.
539.
by ilaksh 2y ago
Yeah, in the last week, I had seen a strong benchmark for chatgpt-4o-latest and tried it for a client's use case. I ended up wasting like 4 days, because after my initial strong test results, in the following days, it gave results that
540.
by ilaksh 2y ago
It's not rambling or word salad. It just tries to include some justification and explicit directions. I have to question the literacy levels of people these days if a page of text is too much to handle.

More ›