News for hackers. Stuff that matters.

supai

A few things:

- We do something similar to OpenRouter which measures the latency of the different providers, to ensure we always get the fastest results

- Users can cancel a single model stream if it's taking too long

- The orchestrator is pretty good at choosing what models for what task. The actual confidence scoring and synthesis at the end is the difficult part that you cannot do naively, however, the orchestrator plays the biggest part in optimizing cost + speed. I've made sure that we don't exceed 25% extra in cost or time in the vast majority of queries, compared to equivalent prompts in ChatGPT/Gemini/etc.

The reason why this is viable IMO is because of the fact that you can run multiple less-intelligent models with lower thinking efforts and beat a single more-intelligent model with a large thinking effort. The thinking effort reduction speeds up the prompt dramatically.

The sequential steps are then:

1. Ensemble RAG 2. Orchestrator 3. Models in parallel 4. Synthesizer

And retries for low-confidence (although that's pretty optimized with selective retries of portions of the answer).

Here's a short demo video: https://youtu.be/DRcns0rRhsg

Show HN: Sup AI, a confidence-weighted ensemble (52.15% on Humanity's Last Exam)

https://sup.ai

Hi HN. I'm Ken, a 20-year-old Stanford CS student. I built Sup AI.I started working on this because no single AI model is right all the time, but their errors don’t strongly correlate. In other words, models often make unique mistakes relative to other models. So I run multiple models in parallel and synthesize the outputs by weighting segments based on confidence. Low entropy in the output...

( Read More... | 26 comments | 2 points | on HN )
Update: Supabase support came through and successfully recovered the database. Our site is back online and no data was lost.

The Supabase dashboard is still experiencing errors and won't load properly, but the underlying data and API are intact.

Thank you to the HN mods for un-flagging this, and to everyone who took a look. Hopefully, we'll receive a post-mortem from Supabase on the exact cause of the disk downgrade issue during the PG update.

Ask HN: Supabase PG upgrade wiped production DB, PITR backups failing

We are currently experiencing a total production outage and severe data loss on Supabase, and we cannot get a response from support. We are hoping someone from their team sees this here.The Timeline of Failure:1. We performed a Postgres version upgrade on our instance. 2. For an unknown reason, this upgrade triggered an unexpected downgrade of our disk size. 3. We ran a standard REINDEX: ...

( Read More... | 1 comment | on HN )