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

Show HN: Gemini Pro 3 imagines the HN front page 10 years from now

dosaygo-studio.github.io

Post reply on HN
Print view View original

761–770 of 1001 posts

Re: Show HN: Gemini Pro 3 imagines the HN front page 10 years from now

#762

Post by schaum »

giancarlostoro wrote: post #86
icyfox wrote: post #50

Exactly half of these HN usernames actually exist. So either there are enough people on HN that follow common conventions for Gemini to guess from a more general distribution, or Gemini has memorized some of the more popular posters. The ones that are missing: - aphyr_bot - bio_hacker - concerned_grandson - cyborg_sec - dang_fan - edge_compute - founder_jane - glasshole2 - monad_lover - muskwatch - net_hacker - oldti…

Before the AI stuff Google had those pop up quick answers when googling. So I googled something like three years ago, saw the answer, realized it was sourced from HN. Clicked the link, and lo and behold, I answered my own question. Look mah! Im on google! So I am not surprised at all that Google crawls HN enough to have it in their LLM. I did chuckle at the 100% Rust Linux kernel. I like Rust, but that felt like a cl…

Every few years I stumble across the same java or mongodb issue. I google for it, find it on stackoverflow, and figure that it was me who wrote that very answer. Always have a good laugh when it happens.

Usually my memory regarding such things is quite well, but this one I keep forgetting, so much so that I don't remember what the issue is actually about xD

Re: Show HN: Gemini Pro 3 imagines the HN front page 10 years from now

#763

Post by Aachen »

josalhor wrote: post #174

"Right to Human Verification" is something I have actually thought about a lot. I want to able to verify my identity against a system. I also want to be able to not do that . So for instance, on Twitter/X, I could verify myself and filter only other verified people / filter those goverments that have validated the identities of the users. I want to be able to do that. But I also want to be able to log in into Twitter…

I tried to place an order on amazon for the first time last week. After creating an account and going through all the steps including paying with a non-refundable payment method (iDeal, similar to bank transfer), I went to bed, only to notice later that it was immediately cancelled "to protect the their systems and their customer" (they thought I was somehow defrauding them by giving them money for a product being offered)

I was welcome to go through human verification, send my passport to some third party, wait three working days, and then they'd review the case... it would never arrive in time for the person's birthday anymore

Human review sounds good but if this is the sort of hoops you have to jump through, you're still screwed if the algorithms hates you because they're everywhere: from trying to use paypal to paying for a translator with stripe to trying to rescue food with "too good to go" using credit cards. Blocked, blocked, blocked. Not that I could get a credit card because the algorithm at the bank didn't like me, but a shared card (where the other person is also responsible for any debt) was okay, so now I need to continuously pay my partner 50% of what I buy with it for myself. You'd think I'm not creditworthy but I never had debts or needed any loan. I don't know why algorithms keep blocking me trying to give other people money

Requiring that any fraud algorithm is fair and transparent would probably go a much longer way than codifying a right to appeal. Appeal processes will either just consist of a human clicking confirm on the algorithm's choice, or have a bunch of extra hoops to jump through. If the algorithm were required to be fair to begin with, you could have them fix the root cause instead by demonstrating that it blocks you for no discernable reason and they need to figure out on their end how to make that not happen

Re: Show HN: Gemini Pro 3 imagines the HN front page 10 years from now

#764

Post by teiferer »

avianlyric wrote: post #628
thomasm6m6 wrote: post #598

Earlier quoted context omitted.

Here it is: https://sw.vtom.net/hn35/news.html I downloaded the original article page, had claude extract the submission info to json, then wrote a script (by hand ;) to run feed each submission title to gemini-3-pro and ask it for an article webpage and then for a random number of comments. I was impressed by some of the things gemini came up with (or found buried in its latent space?). Highlights: "You’re probably…

Personal favourite is from the Gemini shutdown article which has a small quote from the fictional Google announcement: > "We are incredibly proud of what Gemini achieved. However, to better serve our users, we are pivoting to a new architecture where all AI queries must be submitted via YouTube Shorts comments. Existing customers have 48 hours to export their 800TB vector databases to a FAT32 USB drive before the ser…

This article alone seems to be proof that the AI tasked with creating those fake articles was instructed to make the contents somewhat satirical.

Re: Show HN: Gemini Pro 3 imagines the HN front page 10 years from now

#765

Post by jjcob »

throwaway2037 wrote: post #725
danlugo92 wrote: post #693

Earlier quoted context omitted.

What about this one: > The micro-transaction joke hits too close to home. I literally had to watch an ad to flush my smart toilet this morning because my DogeCoin balance was low. And the response...

I am nearly in tears after reading this chain of posts. I have never read anything so funny here on HN. Real question: How do LLMs "know" how to create good humor/satire? Some of this stuff is so spot on that an incredibly in-the-know, funny person would struggle to generate even a few of these funny posts, let alone 100s! Another interesting thing to me: I don't get uncanny valley feelings when I read LLM-generated…

The jokes are not new. If you read Philip K Dick or Douglas Adams there's a lot of satirical predictions of the future that sound quite similar. What's amazing about LLMs is how they manage to almost instantly draw from the distilled human knowledge and come up with something that fits the prompt so well...

Re: Show HN: Gemini Pro 3 imagines the HN front page 10 years from now

#766

Post by zwnow »

hn_throwaway_99 wrote: post #666

This is awesome, but minor quibble with the title - "hallucinates" is the wrong verb here. You specifically asked it to make up a 10-year-in-the-future HN frontpage, and that's exactly what it did. "Hallucinates" means when it randomly makes stuff up but purports it to be the truth. If some one asks me to write a story for a creative writing class, and I did, you wouldn't say I "hallucinated" the story.

If someone asked you, you would know about the context. LLMs are predictors, no matter the context length, they never "know" what they are doing. They simply predict tokens.

Re: Show HN: Gemini Pro 3 imagines the HN front page 10 years from now

#767

Post by thomasm6m6 »

teiferer wrote: post #764
avianlyric wrote: post #628

Earlier quoted context omitted.

Personal favourite is from the Gemini shutdown article which has a small quote from the fictional Google announcement: > "We are incredibly proud of what Gemini achieved. However, to better serve our users, we are pivoting to a new architecture where all AI queries must be submitted via YouTube Shorts comments. Existing customers have 48 hours to export their 800TB vector databases to a FAT32 USB drive before the ser…

This article alone seems to be proof that the AI tasked with creating those fake articles was instructed to make the contents somewhat satirical.

script is here: https://github.com/vtomnet/hn35/blob/main/generate.py

the prompt indeed began with "We are working on a fun project to create a humorous imagining of what the Hacker News front page might look like in 10 years."

Post reply on HN