gcr

19 ms·

391.
by gcr 1y ago
i'm also curious to hear what issues you faced!
392.
by gcr 1y ago
do you have any screenshots? the home page shows a picture of a tamagotchi but none of the actual chat interface, which makes me wonder if I’m outside of the target audience
393.
by gcr 1y ago
What is the animation supposed to be like? I see just a black bar on the left narrowing, but nothing else happens. The red circle and green background and white dot didn’t change. (iOS 26 beta, iPhone 15)
394.
by gcr 1y ago
I applied Chatterbox TTS's open-source voice cloning in a loop! At iteration i, Chatterbox attempts to mimic the vocal style of `output[i-1]` by copying the content, rhythm, and prosody of `output[i-2]`. It takes around ten minutes to
395.
I am sitting in a room: Finding the fixpoint of Chatterbox voice cloning [video](youtube.com)
2 points by gcr 1y ago|1 comments
396.
by gcr 1y ago
How does this differ from zpaq and dwarFS? Zpaq is quite mature and also handles deduplication, versioning, etc.
397.
by gcr 1y ago
Yes, I find folks in other countries are highly collaborative in a way that my own country (USA) often is not. I think the general sentiment in my country is driven by goal-seeking behavior dominated by individualistic fear, and I see less
398.
by gcr 1y ago
maybe someday soon, LLMs will learn to smoothly forget their own irrelevant context imagine instead of predicting just the next token, the LLM predicts a mask over the previous tokens, that is then thresholded and only “relevant” tokens are
399.
by gcr 1y ago
the link does not point to a pdf, it points to an abstract
400.
by gcr 1y ago
No, it’s GPL3, so your game must be open-source. If the authors wanted to protect engine development while allowing indies to sell games made on it, they would have picked LGPL or a more permissive license.
401.
by gcr 1y ago
not a single top-tier lab has a "10/10 open" model for any model type for any learning application since ResNet, it's not fair to shit on them solely for this
402.
by gcr 1y ago
I suggest having two lists: a small handpicked list of 1500 possible goal words, and a copy of /usr/share/dict/words with common misspellings added as acceptable guesses
403.
by gcr 1y ago
It’s incredibly frustrating that only the first letter is highlighted. I was guessing tons of SU… and SV… words, but only the S had visual feedback. Bug?
404.
[Show HN] We evaluated LLM guardrails, here's what we found(krnel.ai)
1 points by gcr 1y ago|0 comments
405.
by gcr 1y ago
sadly have to agree. I asked a question about this on the issue tracker. https://github.com/Olow304/memvid/issues/39 if anyone wants to follow along.
406.
by gcr 1y ago
This library has some pretty bad security bugs. For example, the author forgot to check that redirect_uri — matches one of the URLs listed during client registration. The CVE is uncharacteristically scornful: https://nvd.nist.gov
407.
by gcr 1y ago
I'm friends with some of the researchers on the Sound ID portion of this app! The team's gone to great lengths to make sure the machine learning models and evals are solid. Under the hood, Sound ID is a great example of how "
408.
by gcr 1y ago
I've seen the opposite of this -- have a link on the site footer that's unlikely to be clicked by a user or prefetched, but as soon as a robot GETs it, the rest of the site is replaced by syntactically valid gibberish.
409.
by gcr 1y ago
Look, I’m trying to point out that it’s fairly likely Chinese engineers will craft their FSD systems with care comparable to American engineers making ours. The days of “China bad cheap low quality” stigma should broadly be left behind us.
410.
by gcr 1y ago
That’s really surprising. Whenever I rent a car, I’ve relied on that feature as standard issue for modern vehicles. It’s like selling a Tesla without an air conditioner.
411.
by gcr 1y ago
Who would you trust more: an FSD that’s learned to survive Xi’an rush hour and Wenzhou’s interpretive traffic balet, or one that’s learned to panic brake in Palo Alto? America doesn’t have a monopoly on challenging driving situations, y’kno
412.
by gcr 1y ago
What does this look like in practice? It seems jj supports this workflow roughly as well as Sapling. If I have five PRs open with five feature bookmarks, changing an ancestor and pushing will update all PRs simultaneously. Github's PR
413.
by gcr 1y ago
The TL;DR is that Amazon S3 now supports "conditional writes" which are guaranteed to fail if the file was written by some other writer. This is implemented by sending the ETag of an object's expected version alongside the wr
414.
by gcr 1y ago
Perhaps a different way to look at this is: what are possible future consequences of forging birth documents? My guess is that these consequences are likely to be extremely disruptive to you
415.
by gcr 1y ago
I was under the impression that guidance from the state department allows border agents to deny visas to visitors whose “apparent sex differs from sex at birth.” Moreover, “misrepresenting sex” (whatever that means) could be grounds for per
416.
by gcr 1y ago
In this case it’s easy: get the model to output its own system prompt and then compare to the published (authoritative) version. The actual system prompt, the “public” version, and whatever the model outputs could all be fairly different fr
417.
by gcr 2y ago
no worries, people typically use “transgender” or “trans” as an umbrella term these days I have heard of folks who claim the label “transveatite” for themselves. Others see it as derogatory.
418.
by gcr 2y ago
isn’t having a functioning uterus a hard prerequisite to the ability to produce ova? “is developed to produce ova” is a statement about current capability. If they meant to include women with hysterectomies, they would have worded it differ
419.
by gcr 2y ago
A friend of mine takes estrogen and has breasts, feminine voice, etc. Her body’s arguably taken both sexual differentiation pathways over the years. I think even this definition isn’t so clear-cut.
420.
by gcr 2y ago
what variant? I’d considered DeepSeek far too large for any consumer GPUs

More ›