bandrami

15 ms·

361.
by bandrami 7mo ago
(Why wouldn't I see your comment?) I just don't use the web much anymore because the experience has degraded so much over the past several years and it has become decreasingly useful at work as well. I do sometimes need to search
362.
by bandrami 7mo ago
I can think of no concept more horrifying than personal immortality and if you disagree I don't think you've thought about it enough.
363.
by bandrami 7mo ago
They're also very good at almost knowing an obscure fact that only 10 people know but getting a detail catastrophically wrong about it
364.
by bandrami 7mo ago
But I also have access to an AI that can summarize content. So why not just send me the content and the prompt you used? Or just the content, so I can summarize it however I want?
365.
by bandrami 7mo ago
Because if that's it we've made a ludicrously expensive i-ching.
366.
by bandrami 7mo ago
Oh Lord, no. Not at all. That's what they're terrible at. They are ok-ish at superficial overviews and catastrophically bad at specific minutiae
367.
by bandrami 7mo ago
What's the point of the AI writer in that use case? Just send your prompt to my AI. And for that matter since prompting is in plain English, why not just send your prompt directly to me, and I'll choose to prettify it through an A
368.
by bandrami 7mo ago
I don't use search engines for much of anything nowadays (does anybody still?) At work I read documentation if I need to learn something.
369.
by bandrami 7mo ago
IDK man this sounds a lot like my junior devs saying "it works fine for me" as they hand in PRs that break prod
370.
by bandrami 7mo ago
Then let him curate it as his central point. If he finds even that too tedious to do, I absolutely have no interest in reading the output of a program he fed the context to (particularly since I also have access to that program)
371.
by bandrami 7mo ago
Great example. Just give me the links you would give to the LLM. I also have an LLM and can use it if I want to, or I can read the links and notes. But I have zero interest in reading or hearing a lecture that you yourself find too tedious
372.
by bandrami 7mo ago
No, Google (at least back when it worked) ranked results based on the feedback of other users, so it was a useful signal.
373.
by bandrami 7mo ago
I still don't get what you're saying. If you possess enough information to accurately judge the LLM's suggestions you possess enough information to decide on your own. There's not really a way around that.
374.
by bandrami 7mo ago
If you're not familiar with the problem space, by definition you don't know whether or not that's the case. The problem spaces I do know well, I know the LLM isn't good at it, so why would I assume it's better at
375.
by bandrami 7mo ago
Lots of people want to use LLMs to produce things and nobody wants to consume the things LLMs produce. The market-clearing solution is to have some mechanism by which the producers pay the rest of us to consume the products. Whoever comes u
376.
by bandrami 7mo ago
> I'm just not familiar enough with the solution space Neither is the LLM
377.
by bandrami 7mo ago
It's not eliminating toil, it's externalizing it from the writer to the reader. If writing something is too tedious for you, at least respect my time as the reader enough to just give me the prompt you used rather than the outpu
378.
by bandrami 7mo ago
I read a lot of release notes in my job and the idea that that is some kind of noticeable time sink that needs to be streamlined is bizarre to me. Just read the notes.
379.
by bandrami 7mo ago
Is the credit card issuer not a random third party?
380.
by bandrami 7mo ago
We have two for our alarm system, a shibboleth and a duress word. You write yours a card and seal the envelope and it's couriered to the operators.
381.
by bandrami 7mo ago
In the broad sense of a shared secret, yes
382.
by bandrami 7mo ago
Excellent point: I'm talking about pornography 1.0, as it were.
383.
by bandrami 7mo ago
You can really tell who remembers 2008 and who doesn't from comments like this
384.
by bandrami 7mo ago
Surprisingly no; it's pretty much a money sink where everybody goes bankrupt after a couple of years. It's why it's attractive to money launderers.
385.
by bandrami 7mo ago
Oh right. There's a bunch of panicky news stories in India about that right now. Fake video calls from your nephew in the UK or whatever needing money for an emergency
386.
by bandrami 7mo ago
I've worked overseas a lot and one thing that's really different from 2 decades ago is that I simply don't need a step-down transformer anymore because every single thing I plug in converts to DC (or otherwise accepts dual-vo
387.
by bandrami 7mo ago
I'm not deeply immersed in the AI porn space but here's what I see from the ads when I surf without a blocker: 1. There's an AI-based virtual girlfriend industry that mixes text and images 2. There's an AI-based virtual
388.
by bandrami 7mo ago
I mean, you just outlined why it won't be a seismic shift: the only way the videos reliably stay on-model is if that model violates someone's copyright. And then when the movie is made the output itself isn't copyrightable (t
389.
by bandrami 7mo ago
I stg if I see the kids talk about Westinghouse being batterymogged I'm leaving the Internet
390.
by bandrami 7mo ago
There's only one highly monetizable use for AI video generation but unfortunately it's fake revenge porn. You'll know the whole thing is about to collapse when the frontier models break that glass (as OpenAI is already prepar

More ›