karimf

Well, that's exactly what I'm doing for my side project.

The last time I checked, open source full duplex model was still so far behind GPT-Live. So, I'm super excited about Dessa and I think it's very cool.

But I don't think it'll replace the classic cascade system for now for most use cases. A couple of reasons:

- Intelligence

- Tool calling

- System prompt?

- Language support other than English

This is awesome. Thanks for pushing the audio pareto frontier forward.

Probably far fetched for now, but I think the next big evolution is building the pareto/much cheaper alternative to GPT-Live-1.

The STT/TTS market is quite saturated, while today, there's almost no cheap/open source alternative to GPT-Live-1.

While this is very impressive benchmark-wise, GPT-6 Astra showed us that benchmarks don't always correlate 1:1 to intelligence of a model.

When Astra launched, I think Artifical Analysis showed that it was on par with GPT-5.6 Sol and lower than Opus or something like that? Then, they updated the scoring.

I hope that more open source models, including this model, to be "as good to use" as Astra.

I think a curriculum like this is neat and might help with interviews since you go wide and have a checklist of things that you need to learn.

I'm on a totally different path for learning inference engineering. I self-host a voice AI app that has ~2000 monthly active users on my own GPU box.

This forces me to learn about production serving, KV cache, quantization, inference engine, observability and economics, prefill optimization since I'm optimizing for TTFT instead of decode speed, and many more.

It's fun since every optimization you do directly translate to a better user experience or allow you to serve more users using the same hardware.

This 100%. I was Anthropic-pilled. I had a $200/mo subscription and I only used Anthropic models. I was frustrated by the verbose output and the writing style. I tried ASD-STE-100, it helped a bit, but it's still too verbose for my taste.

Then I tried GPT 5.6 Sol. It's night and day.

I think Anthropic just RL too hard on coding capabilities and never calibrated or benchmarked the writing styles.

Not sure why it's on the front page now, but I highly recommend using llama.cpp for running AI model locally vs using other inference framework, unless you have a very specific requirement.

ggerganov and the team have done a stellar job maintaining the quality while still being fast to implement new models/improvements.

Most people are going under identity crisis right now because of recent LLM advancements. This post is a good example that shows that it's not only happening at the individual level, but also on the company/organization level.

Is it still worth building products or companies that can be one-shotted by AI? Probably not.

One interesting consequence is that this force everyone to be more ambitious and do something bigger that's impossible before.

I hope more people are working on something that can always bring net positive to humanity even if there are hundreds of people working on the same thing, like clean energy.