ilaksh

21 ms·

391.
by ilaksh 1y ago
Will be interesting to see how fast inference ASICs, diffusion LLMs, architectural changes like IBM granite small (when is that coming to OpenRouter?) and slight compromises for pre-generation can speed this up. Also I wonder if eventually
392.
by ilaksh 1y ago
He just named himself in honor of something he remembered fondly.
393.
by ilaksh 1y ago
They did say some packages were more. I bet some are 5%, maybe 10 or 15. Maybe more. Well one example could be llama.cpp . It's critical for them to use every single extension the CPU has move more bits at a time. When I installed it
394.
by ilaksh 1y ago
Anyone else prefer to use web components without bundling? I probably should not admit this, but I have been using Lit Elements with raw JavaScript code. Because I stopped using autocomplete awhile ago. I guess not using TypeScript at this
395.
by ilaksh 1y ago
Maybe we should stop talking about Hetzner before they raise their prices guys.
396.
by ilaksh 1y ago
What LLM do you guys use for fast inference for voice/phone agents? I feel like to get really good latency I need to "cheat" with Cerebras, groq or SambaNova. Haiku 4.5 is very good but still seems to be adding a second of la
397.
by ilaksh 1y ago
Cool. And the other person implies that the queries can search across all rows if necessary? For example if all images have people and the question is which images have the same people in them. Or are you talking about a different project?
398.
by ilaksh 1y ago
Does this use CLIP or something to get embeddings for each image and normal text embeddings for the text fields, and then feed the top N results to a VLM (LLM) to select the best answer(s)? What's the advantage of this over using llama
399.
by ilaksh 1y ago
See also https://livescript.net/
400.
by ilaksh 1y ago
Doesn't seem high enough resolution.
401.
by ilaksh 1y ago
I had one client who made their whole thing about knowledge graphs, which I worked on because I needed money and it was interesting, but I am still a little suspicious that they may have had "knowledgebase" and "knowledge gra
402.
by ilaksh 1y ago
It looks to me like n8n has suddenly taken over to the point where Automation is now almost synonymous with n8n for most people. As someone who builds workflows using my own agent system which is based on checklists, subtasks and tool calli
403.
by ilaksh 1y ago
LLMs have an obvious application of removing the command limitations. https://github.com/UlfarErl/lampgpt "FYI, it is now possible to play all of the Infocom games with a phenomenal parser using LampGPT, with just
404.
by ilaksh 1y ago
"Rules, triggers, and workflows can be embedded directly into payments, making them smarter and adaptable." -- are they smart contracts? What kind of workflows? Does the workflow involve sending messages to agents or making HTTP r
405.
by ilaksh 1y ago
I wonder if there could be some kind of platform where you have to pay a $5 deposit or something to be able to post bugs. If you waste people's time with total nonsense then you lose the $5 and can no longer report. If it's less e
406.
by ilaksh 1y ago
GLTron or Armagetron Advanced?
407.
by ilaksh 1y ago
I think it's actually just an optical illusion. Their FAQ mentions giving you a video editor template for adding lighting and shadows.
408.
by ilaksh 1y ago
Wouldn't that be using vibe coding to fix vibe coding? And isn't the expected timeframe hours, days or weeks rather than milliseconds or seconds like people expect with MCP?
409.
by ilaksh 1y ago
https://youtu.be/ZOTqx9Nz4Ws?si=vwLkKPjgTT2Pn_A0
410.
by ilaksh 1y ago
I have built such a tutoring system.
411.
by ilaksh 1y ago
Please parse that sentence more carefully. I said "labor productivity".
412.
by ilaksh 1y ago
Can you try to read my comment more carefully?
413.
by ilaksh 1y ago
If you look at the actual ranges on his graphs, the conclusion is a little bit exaggerated, but the basic point is valid. Education systems need to have very strong curriculum about AI use and the necessity of just not using it to some degr
414.
by ilaksh 1y ago
Does it need to maintain the 90C or the energy is lost?
415.
by ilaksh 1y ago
The thing that bothers me so much about this post is that it reeks of privilege by treating unemployment as if it were mainly just a bad trigger that you talk to your expensive therapist about once a week. For most people in the real world,
416.
by ilaksh 1y ago
Absolutely. Cryptocurrency should make payment middlemen obsolete. PayPal should get laughed out of the building here. It's incredible that this has any legs at all. This thread should be full of the names of Ethereum wallets. MetaMask
417.
by ilaksh 1y ago
What's the easiest way to install the newest nginx in Ubuntu 24? PPA or something?
418.
by ilaksh 1y ago
https://www.johnayliff.com/games/seedship/index.html?blcOndk...
419.
by ilaksh 1y ago
Maybe they have combined an LLM or something with the speech detection convolution layers or whatever they were doing. Like with JSON schemas constraining the set of available tokens that are used for structured outputs. Except the set of t
420.
by ilaksh 1y ago
I think that XP was the only true agile methodology. Agile just got more and more corrupted over the years through stupidity. Clearly AI programming allows you to quickly close feedback loops. I don't think everything needs a comprehen

More ›