HNHacker News
TopNewBestAskShowJobs

voiceblue

640 karma · joined September 23, 2023

submissionscomments
voiceblue··on Rethinking Text Resizing on Web
I think you missed their point. Most people end up needing reading glasses as they age. I tried to look for a specific number, but the consensus on who eventually needs them seems to be: everyone.
voiceblue··on AI doppelgänger experiment – Part 1: The training
That's common enough that there's a name for it [0]. "Stealing the style" is not so much a problem as commoditization, which may find some measure of defense in terms of trademark (not copyright). In this case, for example, the model is named hollieMengert_v1, and we are referring to the style as "Hollie Mengert's art style".

[0] https://en.wikipedia.org/wiki/Clean_room_design

voiceblue··on AI doppelgänger experiment – Part 1: The training
Related: "Invasive Diffusion: How one unwilling illustrator found herself turned into an AI model" [0].

It's astonishingly simple to train both LoRAs and Embeddings nowadays -- if you combine it with tech like ControlNet, you can more or less have direct control over the illustration as well (see this example I converted to Hollie Megert's style with barely any effort [1] using the model from [0]) -- but FWIW I was playing with the WhatsApp AI stickers recently and human-made illustrations still have a "je ne sais quoi" that sets them apart, vaguely reminiscent of "realistic" animation vs. "squash and stretch". I would not be worried as an artist*, but I would be worried if I am making assets for corporations in Corporate Memphis [2].

[0] https://waxy.org/2022/11/invasive-diffusion-how-one-unwillin...

[1] https://i.imgur.com/ZwtKug5.jpeg

[2] https://en.wikipedia.org/wiki/Corporate_Memphis

* In fact, I think I'd be excited, it can be a great tool, and people who lack artistic sensibility might find it almost as hard to use as to make art themselves. However, I do think there needs to be some kind of protection from having your 'style' commoditized without consent, using unauthorized images of your own art.

voiceblue··on Sam and Greg's response to OpenAI Safety researcher claims
> The reason why everybody thinks OpenAI looks particularly bad is because everybody is watching them

A reason, sure. The reason? Come on. There's a clue in their name about what part of their history is particularly controversial.

voiceblue··on Nvidia CEO says future of coding as a career might already be dead, due to AI
What he actually said (according to TFA anyway, I haven't verified he actually said this):

"It is our job to create computing technology such that nobody has to program. And that the programming language is human, everybody in the world is now a programmer. This is the miracle of artificial intelligence."

How do we get from this to "future of coding as a career is dead"?

voiceblue··on I don't want to spend my one precious life dealing with Google's AI search
I don’t think that comment was in regards to AI, I think it was about the blurb that appears instantaneously from Google’s knowledge engine (e.g. when searching a celebrity’s age).
voiceblue··on [dead]
Looks like there might be a need for a free "shill account name" generator.
voiceblue··on Show HN: Serverless collaborative notion-level note editor using CRDT in GO
Yes. Yjs is one such project: https://github.com/yjs/yjs
voiceblue··on Dear Europe, please wake up
Does the data presented here distinguish between full-time and part-time employment? Or hourly and salaried employment? I wasn’t able to find any information other than “hours worked per person employed”, which seems a rather vague statistic to be used to support any claim.
voiceblue··on DMCA Notice Targeting 'Bypass Paywalls Clean' Isn't the Thing to Get Angry About
Having been a victim of DMCA abuse, sometimes it's just easier to shut shop if you're not feeling it to be worth your time and effort to counterclaim and partake of legal battles. This is especially the case if you never bothered to form a legal entity in the first place (which is an indicator that the product/project was not worth such extra effort).

In this sense I think the prevalence of such DMCA notices has made the internet a bit of a hostile place, just this month we were reading about it here. [0]

Is there any chance these laws will be fixed?

[0] https://news.ycombinator.com/item?id=39934696

voiceblue··on I Lost Faith in Kagi
Do you still have the same complaints when selecting “Verbatim” under search tools on Google?
voiceblue··on I Lost Faith in Kagi
I was also shocked to see this number given how frequently the product is mentioned on here. Makes me wonder what the Venn diagram of HN and Kagi users is like.
voiceblue··on I Lost Faith in Kagi
I don’t have a horse in this race, but from using Google’s AI search previews I am not bullish on the future of search being LLMs even with RAG or MoE (though of course, there is plenty of AI to be had apart from these, and I am bullish on AI in general). Even simple things, such as asking about setting up a Starbucks “franchise” (you can’t), are met with enthusiastic, affirmative, and blatantly incorrect answers.

Anyone who is thinking of putting LLM output directly in front of customers (no HITL) had better think twice about this.

voiceblue··on Code search is hard
The business end of gThanks
voiceblue··on Ask HN: Most efficient way to fine-tune an LLM in 2024?
This thread reminds of a competition I once joined where we were supposed to fine-tune an LLM to fill out trivia answers, and we were expressly disallowed from training on the validation set.

However: we were allowed to pick any base model in a given repo. All of the teams that “won” did so for the same reason: they had all picked the same base model (whereas a majority of teams picked the given default), presumably the one that had at some point been trained on the most favorable data for this particular challenge.

It was quite silly. Had everyone had the same base model we’d have a bit more of an interesting problem (more around NLP and alignment than picking the ‘best’ model).

voiceblue··on Launch HN: Aqua Voice (YC W24) – Voice-driven text editor
Thanks! As I mentioned, I've been looking into this for a while. I'll add it to my list.

Now I have:

    - vetrec.io
    - Abridge
    - Scribeberry
    - Scribematic
    - Notezap
    - Lytte
    - Deepscribe
    - FreedAI
    - s10 AI
    - Nable
    - DeepCura
    - DAX copilot
    - Suki
    - M*Modal
    - Amazon Healthscribe
Slightly different (?) - maybe more human:

    - Overnight Scribe
    - Rev.ai
voiceblue··on Launch HN: Aqua Voice (YC W24) – Voice-driven text editor
From exploring this segment for a while, I believe that dictation software is the "brick" in the hair-on-fire analogy (hence, it provides some relief, but is far from an actual solution). There is a form of water (scribes on retainer) but it is too expensive for all but the most profitable of specialties. The problem to be solved is not "dictation but better," but "take this cognitive load away from doctors, and keep notes accurate." (which is what scribes with experience do.) In a broader sense, the problem to be solved is the American healthcare/insurance system (the reason these notes have to be taken in this way in the first place)...

> Less time spent taking notes == more time with patient.

This can be true in some cases, but from what I understand, industry wide it would end up more like:

Less time spent taking notes == more patients scheduled.

Which is still of value, but fails to solve the original point of a physician's frustration, and possibly makes it worse (assuming the physician is still the one generating, handling and verifying the notes, but with better efficiency).

voiceblue··on Repairing a buggy board arcade machine
Nice work, when you write up a post like this it makes it look easy. The reality is often fraught with frustration, not in the least thanks to the excruciating wait for parts to arrive once you've painstakingly found the next fault in the chain.

In the end though, I think it's all worth it, if only for the feeling of getting it done and over with. I've repaired a handful of such machines in the past and it's remarkable how available the schematics and parts tend to be, there are whole businesses dedicated to selling cap kits and parts for old arcades.

voiceblue··on Game of Life, simulating itself, infinitely zoomable
You might enjoy watching the emergent behavior in the execution of this 256 byte program - a mind is born [0].

[0] https://www.youtube.com/watch?v=sWblpsLZ-O8

voiceblue··on Launch HN: DryMerge (YC W24) – Automate Workflows with Plain English
Is there a reason you're limiting yourself to English? This works perfectly for example:

「毎時、おもしろいミームのリンクを送ってください。」

This also works just fine:

मुझे रोज़ दो चार चुटकुले ईमेल कर देना।

voiceblue··on The Google employees who created transformers
For phrases, Google's TF glossary [0] is a good resource, but it does not cover certain subsets of AI (and more specifically, is mostly focused on TensorFlow).

[0] https://developers.google.com/machine-learning/glossary

voiceblue··on The Google employees who created transformers
> but we didn't see what another group found: that scale in compute with simple algorithms were just better

The bitter lesson [0] strikes again.

[0] http://www.incompleteideas.net/IncIdeas/BitterLesson.html

voiceblue··on Lambda on hard mode: serverless HTTP in Rust
As the other commenter points out, this offering isn't quite comparable to Lambda directly. This ends up comparing apples to oranges here and there, but overall I was able to get a good idea of the choices made and the tradeoffs involved. Nice work! I do have a complaint about the comparison table that shows 'convert HTTP to function calls' as an alternative to load balancers and reverse proxies: as we see later in the article, there is still a load balancer involved, and that table creates a false impression that there isn't.
voiceblue··on Two-Thirds of U.S. Adults Would Rather Wait to Watch Movies on Streaming
I work for a very large company whose name starts with a G, and they recently put on an event for employees to watch a movie. The person next to me was on their phone the whole time (they appeared about halfway through the movie) and even took a call during the movie while remaining in their seat. If it wasn't a corporate event I would have left. As such, I couldn't even say anything since it wasn't like I had paid for the movie and I didn't want bad blood.

I was utterly baffled a working professional at this particular company could behave like this around their co-workers, thankfully no member of the general public has ever done something like that around me.

voiceblue··on Vision Pro: What we got wrong at Oculus that Apple got right
Other headsets are like a DC film and the Vision Pro is like a marvel film before endgame. The differences (that matter in GP’s context) are qualitative rather than quantitative, and a big distinction is the shift from VR gaming, which most incumbents and current VR fans won’t “get” nearly as fast as the public will (once things get going).

Somewhat ironically, I have no interest in showing people my AVP, I use it for hours everyday but showing it to people feels weird in the same way showing your neighbor your workstation and monitors feels weird. (Part of that is, of course, that I’m not going to give them my laptop to run Mac virtual display, which means they won’t fully “get” it anyhow.)

In contrast the quest devices are practically begging to be shown off but lie in disuse a majority of the time, especially if you don’t have time for gaming.

voiceblue··on Google's morale crisis is about to get worse
The first and last sentence here seem to be at odds. Am I missing something? GSUs?
voiceblue··on Learning to be managed
If you want to be promoted, find a manager who has a track record of getting people promoted and transfer to them, then do the above. Many managers (and to be fair, environments and factors outside of their control) are awful in ways that boggle the mind, and no (sane) amount of effort will help you in that case.

The other method is to quit, if you are actually better than your level but your company fails to acknowledge it for whatever reason.

voiceblue··on "Coding" was never the source of value
AI “feels” more efficient for programming since it spits out code in a jiffy, but once I am done with a session of coding I really have to wonder whether it wouldn’t have just been faster to look up APIs and online discussions rather than repeatedly trying to fix what the LLM was doing or attempting to force it down a given path if you know roughly what you want. (Plus half the time it’s not clear why a method it wants to use isn’t there - not imported, or doesn’t exist and never existed, or defined elsewhere in whatever context used to be around this in training?)

It’s almost like using StackOverflow but a toddler is in charge of inputting your queries and only tells you a fraction of what he finds.

voiceblue··on Why Some of Apple's Biggest Fans Are Returning Their Vision Pros
Yes, the iPad app works and there is a VP app in beta. [0]

[0] https://testflight.apple.com/join/4eE59dyH

voiceblue··on Why Some of Apple's Biggest Fans Are Returning Their Vision Pros
It depends on what you’re trying to do. This is part of why I suggest trying it for yourself, it’s very hard to get a good sense from what others can say.

As for me, I wouldn’t replace my desktop with this personally (albeit I might get rid of my monitor) but it would very likely replace my MacBook Air at some point.

← PreviousPage 3 of 6Next →