Hacker Newsnew | past | comments | ask | show | jobs | submit | Tsarp's commentslogin


ASR has already proved its usefulness. Dictation tools are a prime example. Ever since whisper came out, usefulness for AST models running locally suddenly became a thing. Opened up soo many variants

https://superwhisper.com

https://carelesswhisper.app

https://macwhisper.com


lanceDB and underlying lance works well as append only right? Wont you have to do index maintenance often for a chat access pattern?

Good question. This is targeted more at smaller chat server use cases that put the admins in the drivers seat. Compaction operations and rebuilding indexes can happen during a maintenance window.

I plan on scaling up the number of messages the bot system generates to see where the breaking point is.


"LLMs themselves are a graph database with probabilistic edge traversal" whaat?

Do you have any good demos to showcase where graph DBs clearly have an advantage? Its mostly just toy made demos.

vector embeddings on the other hand no matter how limited clearly have proven themselves useful beyond youtube/linkedin thought leader demos.


It comes from people who develop LLMs. Anthropic and Google. References below.

My other favorite quote: transformers are GNNs which won the hardware lottery.

Longer form at blog.ladybugmem.ai

You want to believe that everything probabilistic has more value and determinism doesn't? Or that the world is made up of tabular data? You have a lot of company.

The other side of the argument I believe has a lot of money.

https://www.anthropic.com/research/mapping-mind-language-mod...

https://research.google/blog/patchscopes-a-unifying-framewor...


Not sure how that was the take away from both the posts above.

I read the blog post and your website but unfortunately didnt help change my perspective.

Thanks for the share


Did it start around the covid/ WFH time? There are a few theories

1. Airpods or ANC

2. WFH ->. Less movement -> stiff muscles around neck and head -> head trasnfer frequencies changing

3. Covid vaccine


I know some people blame the covid vaccine. I had tinnitus before the vaccine and it got louder when I took the vaccine and then got quieter later. But it gets louder with other medications too. I suspect anything that causes inflammation can increase tinnitus symptoms, and the covid vaccine does temporarily increase inflammation. This could easily push someone who hasn’t noticed their developing tinnitus over the edge and suddenly they notice it and associate it with the vaccine. What they don’t know is that they might have noticed their tinnitus 2 months later if they hadn’t taken the vaccine. Statistically, I would expect there to be lots of people, like maybe as much as one or two percent of the population (which amounts to a few million people in the US) who might legitimately associate their tinnitus with the covid vaccine, even if the vaccine actually had nothing to do with it.

The same explanation goes for ANC - when you cut out all the noise, suddenly it’s way easier to notice the tinnitus you already had.

There might be something with the neck stiffness idea. I do get the feeling my tinnitus lessens when I’m using cervical traction and doing neck stretching regularly.


See this is why we can't have cheap tinfoil.


1. Without noise you become more aware of your tinnitus.

2. WFH -> Less movement -> Decreased blood flow can contribute to the onset of tinnitus.

Long exposure to high volumes causes hearing damage. Many people set volume on headsets too high to hear better.

3. Many people are diagnosed with tinnitus every day, and some are bound to have it discovered after a vaccine shot. In the same way, some people will have tinnitus discovered after COVID. That doesn’t yet prove causation.


what does this have to do with Thinking Machines? Just clickbait?


Making local running dictations transcribe faster than cloud tools even for longer dictations. <2.0 seconds for any length of dictation.

https://carelesswhisper.app


Handy is missing from the comparison page ;)


Voice interfaces. Currently this dictation app https://carelesswhisper.app

Managed to make long dictations even >10mins appear in < 2seconds by pushing what is possible with current STT models.

All processing done locally with 0 network calls.


If you were to classify how you landed the 170+ clients into a few buckets what would the top channels be?


Upwork, majority of them (i'm top ~40 freelancer there and at one point was top ~10). The rest, non-repeatable, non-systematic randos.


Sounds like you solved the trust issue with your upwork profile reputation.


Making local running dictations transcribe faster than cloud tools even for longer dictations. Yes its possible.

https://carelesswhisper.app/blog/latency-demo


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: