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

Agreed. I am open to the possibility of the bubble bursting or whatever, but this piece is like 3,000 words and cites everything as evidence the sky is falling. It's just as bad as the pro-AI grifters, just in the other direction.

Does the truth normally lie somewhere in the middle of it all?


At the /. times, there was somebody there with the best signature line. It was something like:

"Some people say the Sun sets at East, other people say it sets at West. The truth, of course, is certainly on the middle."


>Does the truth normally lie somewhere in the middle of it all?

Usually does when you decide what constitutes extreme.


Probably. Although I feel more inclined to forgive Ed in this case because it's sort of fighting fire with fire, the insanely hyperbolic and obscenely misleading drivel that's coming out of the most ardent AI boosters is continually unchallenged in the public eye. In a world where we had a more realistic view of AI/ML/LLMs, the limits to its capabilities, and the negative externalities of its widespread adoption in places where it quite frankly does not belong, then I'd be more critical of the Chicken Little sort of writing style

Tell us one or two, then!

The (fun?) one that stands out the most were the Pittsburgh Angels (2013 probably) and they literally have a long round table that they dine at in a closed off room in this country club. They take like one or two pitches a month and tell you that they generally do 300-500k checks.

I was invited and drove 3 hours. They then put you in a giant dining room with all the country club members eating dinner, with your own table by yourself. They tell you to order whatever you want, and wait there, and when the group is ready you'll be summoned to dance.

Anyway, whenever they are ready which was like an hour and a half, they call you in and you have to be ready.

While they loudly eat dessert you are supposed to walk around this 20' long table pitching them whatever you're doing and then you have to take questions from them, while still walking around a table. No presentation place, no place to sit, no real way to talk with everyone, you're always talking over someone's head.

Then they say thank you and on the way out you get a bill for whatever you ate.

That was one of the "nice" ones. There's tons of bullshit networks like TiE, Keiretsu forum, YoungPresidentsOrganization shit like that are all social clubs for wannabe oligarchs.

I think every Chinese backed firm on sand hill road at the time was a front for Chinese billionaires money laundering, I'm sure I have the giant spreadsheet of funds I talked to from 2016 that later turned out to be almost exactly that like Rothenberg Ventures.

The bad ones are LP's sexually assaulting people and being told to ignore it or that it's no big deal (I didn't ignore it and people got fired). Or when LPs push CEOs to fire more people than they need to, or to raise money through some shady vehicle to get it off the books. Not to mention seeing too many people that later were in the Panama/Epstein files.

Literally horrific


They tell you to order whatever you want, then surprise you with a bill at the end?

I was also baffled

Not a great comparison; those apps have serious network effects.

Already on the front page:

The newest Instagram “exploit” is the goofiest I've seen

https://news.ycombinator.com/item?id=48359102 - 180 comments


I know that HN requests that we don't editorialize the titles, but I feel like the article title for this thread better expresses what's happened at a glance than the "goofy exploit" article.

When the title is too editorial HN staff will fix the title to be more accurate. But probably only if you email hn@ycombinator.com and ask for it! Without your help surfacing issues, there is too much volume day in and day out to keep up.


Don't forget about Contrastive Negation:

> Contrastive negation is a rhetorical structure that denies a specific idea in the first half of a sentence and asserts an alternative in the second half.

> It typically follows an "It’s not X, it’s Y" or "not just X, but Y" formula.

Wikipedia also has a great resource which covers many of the common LLM patterns: https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing


As someone who’s writing gets flagged as AI a lot, I appreciate the opening disclaimer section.

And I especially appreciate the idea that these patterns aren’t the problem. Don’t just fix the patterns. The problem is they point to hastily or carelessly written content.


That's in the post


How long until LLMs are taught to avoid the patterns in that wiki page?


Your numbers are a little off but the point remains- 2PB is nothing, not newsworthy imo. What’s special about this?


What's special about it is not the flash but training an LLM based on the content, much of which is still in copyright and which the library has restrictions on how they are allowed to use (irrespective of the legal position of training on it) and which required an agreement with the copyright holders.


This is super useful. Most of the time I go to run a model off Hugging Face on my 64GB MBP I run into issues where I drastically overestimated what it could do. :>



Showing up is a trap for Anna - who doesn't have 5 billion dollars to settle.


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

Search: