6,919 karma · joined November 7, 2014
The only reason why other countries sent their soldiers to stupid wars in Afghanistan and Irak that made no sense then and make no sense now was because they were honoring their commitments and saying OK, we send people to die but if the worst come and Russia attacks us, The USA will do their part. Now Trump openly says all this was for nothing. All trust in the USA is gone as nobody knows what Trump does when he wakes up next day. So we quietly work on making the best of the current situation.
An API call to GPT4o works quite well (it basically handles both transcription and diarization), but I wanted a local model.
Whisper is really good for 1 person speaking. With more people you get repetitions. Qwen and other open multimodal models gives subpar results.
I tried multipass approach, with the first one identifying the language and chunking and the next one the actual transcription, but this tended to miss a lot of content.
I'm going to give canary-1b-v2 a try next weekend. But it looks like in spite of enormous development in other areas, speech recognition stalled since Whisper's release (more than 3 years already?).
For a few months I used Gemini Pro, there was a period when it was better than OpenAI's model but they did something and now it's worse even though it answers faster so I cancelled my Google One subscription.
I tried Claude Code over a few weekends, it definitely can do tiny projects quickly but I work in an industry where I need to understand every line of code and basically own my projects so it's not useful at all. Also, doing anything remotely complex involves so many twists I find the net benefit negative. Also because the normal side-effects of doing something is learning, and here I feel like my skills devolve.
I also occasionally use Cerebras for quick queries, it's ultrafast.
I also do a lot of ML so use Vast.ai, Simplepod, Runpod and others - sometimes I rent GPUs for a weekend, sometimes for a couple of months, I'm very happy with the results.
Why? I mean, if this was another president and another scenario, I'd agree. But I can very well imagine Trump taking Greenland by force and at the same time leaving all American units in Europe as they are.
This is what everybody hopes for, myself included.
I mean good for them, but it would be nice if the same happened to e.g. Astral (cf.).
And on the foreign front, well, the trust has been broken and I don't think it can be repaired easily. The remaining NATO allies are very stable, they understand mutual respect and collaboration is the foundation of their survival, whereas attacking each other breaks trust completely. Even if a new fantastic president is chosen that understands his huge responsibility both for Americans and the world, other countries learned their lesson hard and understand there is no guarantee in a few years the USA become their enemy again.
Correct me if I'm wrong but these and other Linux patches were always about soft real time and Linux never had hard real time capability because of its architecture.
I know nothing about the subject, just wanted to say your comment is exactly one I'd prefer to filter out - it adds nothing to the discussion, I haven't learned anything from it and reading it makes me feel like wasting my precious time on Reddit rather than on one of the very few remaining quality forums like HN.
(Yes, propaganda is present in all countries, but if you eliminate all opposing voices, the pendulum dangerously sweeps towards one side.)
As the self-contradictory saying goes, "all generalizations are false"; nevertheless, some Chinese I met are definitely conditioned by Chinese propaganda in a way that doesn't stand closer scrutiny. Very nice, well-educated people, and touch the subject of the Dalai Lama and see the fury unfold.
I can offer a slightly different perspective. I remember Microsoft from the 90s and early 2000s. And while technical details differ, their attitude towards users didn't change that much.
The problem with this approach is that in 30 minutes, an agent is able to produce a massive amount of stuff. Reviewing all this is a nightmare, in the sense that on the surface it seems fine and it often works, until it doesn't. The bugs introduced are often subtle and their effects manifest later, if ever.
So, for stuff that matters (to me), I prefer not to use agents at all.
Maybe things will change in a year, or 5, or 10. I will be giving it a try. but for the moment it's just not worth it, and the upside-down workflow it pushes on me is just making me tired and lose satisfaction from doing my job.