Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Amazon has taken down the Sad Bastard Cookbook

wandering.shop

Post reply on HN
Print view View original

181–190 of 202 posts

Re: Amazon has taken down the Sad Bastard Cookbook

#181

Post by squeaky-clean »

fsckboy wrote: post #135
borlanco wrote: post #118

Earlier quoted context omitted.

I would say yes. They try hard to make cooks laugh. This is one of their recipes: APOCALYPSE RAMEN This is basically like regular ramen soup but with three critical differences: - It is intensely chaotic, which is to say that you can pretty much toss anything into it and say you did it on purpose. - It's slightly healthier, if your depression has lifted enough that you want to eat something other than delicious, deli…

put boiling water in a jar and then shake it up? Ouch! when you agitate hot water, it evaporates a lot faster, bringing the pressure in the jar to great heights. If your seal is not good, or if you open the jar scalding water can come shooting out onto your hands. a similar thing happens when you put boiling water in a blender and turn the blender on (your blender seal is not good) explosive results.

The recipe literally says to wait until the water is no longer boiling hot though. Anyone can harm themselves cooking if they don't follow the actual recipe steps.

Re: Amazon has taken down the Sad Bastard Cookbook

#182

Post by SoftTalker »

DennisP wrote: post #152
cdrini wrote: post #31

My guess would be this is an overzealous AI-generated content detector. Amazon has been flooded with low quality AI generated books in the past ~year, and has been removing a lot of it. My guess would be they're still tweaking their algorithms as folks generating books change their tactics. Hopefully a human will respond soon!

Here is Amazon's policy on AI-generated content: > We require you to inform us of AI-generated content (text, images, or translations) when you publish a new book or make edits to and republish an existing book through KDP. AI-generated images include cover and interior images and artwork. You are not required to disclose AI-assisted content. https://kdp.amazon.com/en_US/help/topic/G200672390 I personally know someon…

WTF is the difference between AI-generated and AI-assisted. Seems like a very murky distinction.

Re: Amazon has taken down the Sad Bastard Cookbook

#183

Post by tomrod »

downut wrote: post #109
Kye wrote: post #95

Earlier quoted context omitted.

I think I would miss PINES if I ever left Georgia. The nearest library is tiny, but I can have any of 10 million books delivered to it. And if the system doesn't have it, they appreciate and respond to requests.

We moved from AZ to suburban Atlanta last year, and were familiar with interlibrary loan for rural libraries in AZ. But wow! PINES is sensational. We get books and DVDs, even some BluRays from all over GA. Been watching The Wire again, it's great. And the inventory isn't censored either. I started a reread of Gravity's Rainbow and that came from Carrolton (college town). Certain gory/raunchy murder mystery series com…

We did the same -- Sequoyah regional library has been really cool!

Re: Amazon has taken down the Sad Bastard Cookbook

#184

Post by ChrisMarshallNY »

SoftTalker wrote: post #182
DennisP wrote: post #152

Earlier quoted context omitted.

Here is Amazon's policy on AI-generated content: > We require you to inform us of AI-generated content (text, images, or translations) when you publish a new book or make edits to and republish an existing book through KDP. AI-generated images include cover and interior images and artwork. You are not required to disclose AI-assisted content. https://kdp.amazon.com/en_US/help/topic/G200672390 I personally know someon…

WTF is the difference between AI-generated and AI-assisted. Seems like a very murky distinction.

I'll give you an example from my recent work.

We just released a new app onto the Apple App Store.

In order to do this, we need to submit a bunch of screenshots, preview videos, and text.

The images and videos, we hand-crafted. A lot of folks automate this, but I think that it's better to hand-craft, if possible, because we need to ensure that specific contexts and whatnot are reflected, and that user privacy is respected.

For the text, we developed a fairly dry blurb, and then I asked ChatGPT4 to rewrite.

I hit "refresh" a few times, and wrote up a text document with the results, and shared that with the team.

We then picked the parts we wanted, and wrote some of our own.

None of the ChatGPT-generated summaries was sufficient, on its own merits. We had to cut-and-paste (sometimes changing the order, and changing to make sure that there was consistency), and add our own content, as well.

Re: Amazon has taken down the Sad Bastard Cookbook

#186

Post by mike_d »

jasode wrote: post #125
cdrini wrote: post #31

My guess would be this is an overzealous AI-generated content detector. Amazon has been flooded with low quality AI generated books in the past ~year, and has been removing a lot of it. My guess would be they're still tweaking their algorithms as folks generating books change their tactics. Hopefully a human will respond soon!

>My guess would be this is an overzealous AI-generated content detector. Maybe not. If you go by the screenshot of the email from Amazon in the Mastodon link, the book was removed for violating "Community Guidelines" instead of the actual content of the book itself. This is the "Community Guidelines" url in that email: https://www.amazon.com/gp/help/customer/display.html?nodeId=... So taking Amazon's reason at face v…

> So taking Amazon's reason at face value

Having dealt directly with managing abuse and handling account terminations, the reason you are told is never the actual reason.

As annoying as it is for the 1% of false positives to not get a straight answer, the net effect is the 99% of bad actors don't know what they did to get booted - which means they can't adapt as easily.

Re: Amazon has taken down the Sad Bastard Cookbook

#187

Post by wegfawefgawefg »

Hasu wrote: post #143
wegfawefgawefg wrote: post #124

Earlier quoted context omitted.

I think its totally the other way around actually. It looks blunt from afar, but when you use it on specific fiddly things, it does a great job. Domain experts get the most out of it. The average person really can't use it well. I do agree with you that output on non domain specific tasks does not stand up to scrutiny. However, I often have to remind people that gpt4, and chatgpt in particular are not really llm's fu…

> Domain experts get the most out of it. The average person really can't use it well. This is the exact opposite of what research has shown so far. Top performers show productivity _decreases_ with LLM assistance, lower performers and less experienced users show a modest improvement. [0] [0] https://www.nber.org/papers/w31161

Customer support agents arent really what i meant by... domain expert...

Re: Amazon has taken down the Sad Bastard Cookbook

#188

Post by wegfawefgawefg »

Kye wrote: post #147
Hasu wrote: post #143

Earlier quoted context omitted.

> Domain experts get the most out of it. The average person really can't use it well. This is the exact opposite of what research has shown so far. Top performers show productivity _decreases_ with LLM assistance, lower performers and less experienced users show a modest improvement. [0] [0] https://www.nber.org/papers/w31161

LLM output reminds me of the essays I had to write in college. They tick all the boxes in a curriculum and demonstrate a basic grasp on syntax, but they're hot garbage. Your first encounter with an editor as a professional or advanced hobbyist burns that away and reveals the foundation of your true form as a writer. LLMs are in desperate need of something that bridges that gap if they're to do anything more than help…

No they only write that way because of rlhf. gpt3 was extremely "creative".

Re: Amazon has taken down the Sad Bastard Cookbook

#189

Post by wegfawefgawefg »

hombre_fatal wrote: post #146
Hasu wrote: post #143

Earlier quoted context omitted.

> Domain experts get the most out of it. The average person really can't use it well. This is the exact opposite of what research has shown so far. Top performers show productivity _decreases_ with LLM assistance, lower performers and less experienced users show a modest improvement. [0] [0] https://www.nber.org/papers/w31161

I reckon the midwit meme applies here. Noobs think AI is amazing. Midwits who never bothered to try to leverage it or can't evaluate it or just regurgitate what they hear think it's shit. I wager these people don't create much so they don't have a domain to really apply it, so they rely on others to report how good/bad/useful/shit it is. Then there are creators or tinkerers who've leveraged AI to do obviously time-sa…

What you can do with a tool is mostly a reflection of yourself.

Re: Amazon has taken down the Sad Bastard Cookbook

#190

Post by lwhi »

malermeister wrote: post #134
akie wrote: post #55

Earlier quoted context omitted.

Overly aggressive denials of simple requests ("please use another word") just make me wonder why you are so upset about it? Blacklist/whitelist DO reinforce the notion that black = bad and white = good, so what's the deal with using a different word? I truly don't understand why people would get upset by simple requests like that.

The term blacklist comes from the term black book, which originated in the 1400s as a reference to "a list of people who had committed crimes or fallen out of favor with leaders". It is called that because King Henry VIII used a literal black book. Not everything is about skin color. Not everything is about American racial dynamics. Stop projecting your issues on everything.

The black book was basically an accounting book.

Stop twisting history to suit a narrative you'd like to be true.

Post reply on HN