Browse Comments — Raw (as collected)

A great repo, solidifying concepts by building real-time production systems.

Everything is okay but you need to post a image for this post?? Seems for reach?

The #1 most used model on OpenRouter right now costs $0.07 per million tokens (Hy3 preview). The #2 costs $5.00 (Claude Opus 4.7). Wild gap. check out https://www.gawk.dev for more upto date information

So something like Qwen, but using significantly more specialists. Is it useable for software dev tasks today?

This is the kind of resource that can genuinely shorten the AI learning curve. Practical, production-ready projects always teach more than endless theory. Huge respect to Abhishek Veeramalla and Oracle for making advanced AI concepts more accessible to the community

Mike Pappas ensemble architectures may indeed outperform brute-force scaling, but don't you think that orchestration complexity, latency, and consistency across specialized models will be the real test at production scale?

Is it on hugging face ?

Haha

Is Deepaseek available in the US?

See Michael Lentz - told ya!

The voice ≠ transcript point is the strongest part of this ; curious how the orchestrator handles cases where the wrong sub model gets picked .. does aggregation recover, or do errors compound? :/

Interesting

Is there a simple way to try it out?

my favorite comment in this piece: “Why should we be paying for this infrastructure? Why should we be paying for their power bills?” That's a great question. The benefits of AI to the general populace must be commensurate with the cost for it to survive. Spell-checking an email or asking a question to which you receive an unsteady, hallucinogenic response is not worth it for the majority of people.

I keep getting this advertising for some reason. I looked into it and their 1% is not OpenAI's 100%. It's not like the general purpose LLM's that we know.

What stands out here is less the headline cost comparison and more the architectural philosophy shift: specialized systems collaborating dynamically rather than one increasingly massive generalist model attempting to do everything at once. If the efficiency and performance claims continue holding up at scale, ensemble-based approaches could end up influencing far more than voice AI alone.

it's a year old...

Kristina Tanasichuk do you find it obsolete in any way? What do you believe has changed in the meantime?

Please see latest Marex note: https://www.marex.com/news/2026/04/the-great-ai-infrastructure-buildout-impact-on-power-and-commodity-markets

You forgot to add Hyperlambda.dev The Best Solution for building AI Agents with every generated code said to be 100% mathematically correct. Being the only LLM AI Agent in the world with the most accurate fine-tuning model, it said to come at a fractional cost of 0.000001% of Claude AI. We should be expecting a tsunami of projects hitting the market just because a model got it right 👍