smallerfish

7 ms·

61.
by smallerfish 8mo ago
> "content" Sorry, but we're talking about models as content now? There's almost always a better word than "content" if you're describing something that's in tech or online.
62.
by smallerfish 8mo ago
Yeah. There's nothing wrong with using AI to draft an article, but ffs have some taste and edit it significantly once the draft is written.
63.
by smallerfish 8mo ago
What is the traffic flow rate in an intersection with a 4 way stop? For single lane, since only one vehicle can be in the intersection at once, and probably takes _at least_ 5 seconds to start from stopped and cross the intersection, I&#x27
64.
by smallerfish 8mo ago
The skin is definitely much better, but a higher than "recommended" dose is definitely (anecdata) effective at bringing up and maintaining the measureable Vitamin D3 level in your blood if you are under the recommended range. It&#
65.
by smallerfish 8mo ago
To my understanding Vitamin D is regularly underdosed. Several points: 1) There are lots of studies that correlate Vitamin D production with sunlight exposure. For example, https://pubmed.ncbi.nlm.nih.gov/20398766/ thi
66.
by smallerfish 9mo ago
> But none of them predicted my Covid infection days in advance. Oura did. It actually warned you, or retrospectively looking at the metrics you could see that there was a pattern in advance of symptoms? (If the latter, same here with my
67.
by smallerfish 9mo ago
> I don’t have to do boilerplate, generally. I meant this more in the sense of there is nothing new under the sun, and that LLMs have been trained on essentially everything that's available online "under the sun". Sure, th
68.
by smallerfish 9mo ago
> Most of the tests are BS too. Only if you don't constrain the tests. If you use agents adversarially in generating test cases, tests and review of results, you can get robust and tight test cases. Unless you're in research, m
69.
by smallerfish 9mo ago
I've been working on a TUI to make bubblewrap more convenient to use: https://github.com/reubenfirmin/bubblewrap-tui I'm working on targeting both the curl|bash pattern and coding agents with this (via smart
70.
Show HN: Bubblewrap TUI – makes bubblewrap's command lines less terrifying(github.com)
3 points by smallerfish 9mo ago|1 comments
71.
by smallerfish 9mo ago
Bubblewrap is a low-level Linux sandboxing tool that uses namespaces, seccomp, and capability dropping to isolate untrusted applications. I wrote this TUI because I find bubblewrap's necessarily long command lines hard to grapple with,
72.
by smallerfish 9mo ago
Hopefully the browsers will improve their built in TTS soon. It's still pretty unusable unless you really need it.
73.
by smallerfish 9mo ago
Bear in mind that some portion of this could be human directed research. I'm doing a research project right now with 1000 things that I'm building profiles on; to build a full profile requires an agent to do somewhere around 100 d
74.
by smallerfish 9mo ago
Sure you can. DJI is banned though.
75.
by smallerfish 9mo ago
What's your experience with the competition? I have never tried gemini in sheets, but don't you think google will get there soon if they're not already?
76.
by smallerfish 9mo ago
This is the most valid criticism. Theoretically in several years we may be able to run Opus quality coding models locally. If that doesn't happen then yes, it becomes a pay to play profession - which is not great.
77.
by smallerfish 9mo ago
No doubt AI slop is a problem. Writing well with AIs is a skill -- there are lots of people who are uncritically just copy/pasting whatever the AI produced on first draft onto the web. But I'd argue that's a "content&quo
78.
by smallerfish 9mo ago
You're claiming 20% of LLM responses are hallucinations, today?
79.
by smallerfish 9mo ago
> And they get confident lies back 2/10 times That _might_ have been true in 2023.
80.
Claude devs complain about surprise limits, Anthropic blames expiring bonus(theregister.com)
6 points by smallerfish 9mo ago|0 comments
81.
by smallerfish 10mo ago
The biggest problem is that when this company goes out of business in 5 years that it'll become a paperweight. I'm still bitter at Logitech for screwing up Squeezebox.
82.
by smallerfish 10mo ago
A video demo would be useful. I can't really tell how much the application is doing from the screenshots. Is it a tool with some smart guidance, or is it doing deep magic?
83.
by smallerfish 10mo ago
Somewhat ironically, this article has significant amounts of AI writing in it. (I've done a lot of AI writing in my own sites, and have been learning how to smother "the voice". This article doesn't do a good job of smot
84.
by smallerfish 10mo ago
Correct, but not the one on the site.
85.
by smallerfish 10mo ago
I like it, but the switch from metric to inches is confusing, and I think introduces a bug - there's no way a sea snail is 5-6 neurons high.
86.
by smallerfish 10mo ago
Google Translate doesn't hold a candle to LLMs at translating between even common languages.
87.
by smallerfish 10mo ago
Piracy is seeing a big uptick because streaming increasingly sucks. 10+ years ago before studios started chasing their own streaming platforms, and Netflix was the only game in town, it was an excellent deal. $10ish, as opposed to $50+ for
88.
by smallerfish 10mo ago
Correct, but the current market is not working. 15+ streaming services is terrible for consumers. Catalogs are compromised. Bigger services can push prices up because they have more stuff. Clearly if there are too few players then there&#x2
89.
by smallerfish 10mo ago
I'm a fan. Injecting a huge catalog into Netflix is a win for consumers who want just one subscription. And injecting studio talent into Netflix (assuming the merge gives WB creatives influence) can only help. HBO's tech sucks. Ap
90.
by smallerfish 11mo ago
Ahhh, TBH I didn't look on the right. I dug through the menu on the left, thinking the right hand bar (which has the rotated labels) was all getting-started/docs related things. In my defense I have a fairly wide monitor and tend

More ›