July 23, 2026
Too many bots in the kitchen?
Show HN: Echo – Fable-level results at 1/3 the cost using open-weight models
AI mashup claims top-tier answers for cheap, but commenters are already side-eyeing the signup wall
TLDR: Echo says it can mix several AI models together and get near-top results for about a third of the cost. Commenters are split between **“this could be the future”** and **“show me the proof before the signup wall and privacy drama.”**
A new project called Echo is pitching a very simple dream: instead of forcing one artificial intelligence bot to do everything, let a whole group of them team up and split the work. Its creator says that, in testing, this AI "group chat" matched the results of a stronger rival system called Fable while costing only about one-third as much to run. That is the kind of claim that should make the internet lean in. And it did — but mostly to squint.
The comments quickly turned into a classic tech-food-fight. Skeptics were not impressed by the first look, with one person dragging the launch for having no easy public proof up front, an AI-made promo video, and a signup page that felt more like a velvet rope than a demo. Another user hit the privacy panic button, complaining there was no simple sign-in, no free try-before-you-buy, and terms that allow training on user data. In other words: nice idea, but why does it already feel like homework?
Still, not everyone came to throw tomatoes. A few commenters said the concept has real potential, comparing it to systems that quietly route your question to different bots behind the scenes. One even suggested this might be where the industry is headed: not one genius AI, but a messy little talent show of specialists. The funniest line of the thread, though, compared replacing one model with many to turning every outage into “a murder mystery.” Brutal, memorable, and very much the mood.
Key Points
- •Echo is an experimental system that routes tasks across multiple open-weight models instead of using a single model for all requests.
- •The project originated from evaluations showing a hypothetical best-selection-and-combination setup could outperform any individual model in the pool.
- •Echo dynamically decides computation budget, participating models, and output-combination strategy for each request.
- •On the author’s first evaluation mix, Echo reportedly beat the best individual model in its pool and matched Fable’s aggregate result at about one-third of the inference cost.
- •The project currently includes a chat interface, an OpenAI-compatible API, and published material on methodology, model results, costs, and limitations.