staticman2

What the author provided is not necessarily the same as what the software forwarded to the model, especially if some sort of "recall" feature is being used.

I asked Gemini and Gemini thinks even knowing Ken's uncle is a junior Ken's father is more likely to be named "John" or "James".

If Gemini is correct ChatGPT is dumb and simply got lucky.

It'll likely start behaving differently if you respond by explaining why you found it's response offensive and condescending. The models tend to be pretty flexible in how they adapt to user preference if you call them out.

It's not even a real Voltaire quote :)

"We are all stochastic parrots."

Is there some good faith meaning of this slogan that I'm missing?

Presumably you don't mean "Humans and LLMs are functionally identical."

What does this slogan mean to you?

It would be pretty trivial to paraphrase the output wouldn't it?

>think that a bunch of radical feminists flipped over his car

Like the cartoon character She-Hulk? That's a funny image.

My stated opinion is anyone who comes to an AI conversation and says "I can't tell the difference between organisms and computers" or some variation thereof does in fact have no trouble in practice distinguishing between between their child/ mom/ dad/ BFF and ChatGPT as is in fact questioning from a position of bad faith.

"There is nothing inherently wrong with training machines on existing data..." doesn't really conflate a machine with an organism and isn't what I'm talking about.

If you instead had written "I can read the Cat in the Hat to teach my kid to read why can't I use it to train an LLM?"

Then I do think you would be asking with a certain degree of bad faith, you are perfectly capable of distinguishing those two things, in practice, in your everyday life. You do not in fact see them as equivilent.

Your rhetorical choice to be unable to tell the difference would be performative.

You seem to think I'm arguing copyright policy. I really am discussing rhetoric.

Nintendo Switch had an efuse system to prevent downgrading to earlier less secure versions of the OS but I don't remembering hearing that Nintendo was actively bricking anyone's Switch.

They would ban online use but I believe it would still work offline.

Anyone writing "humans can learn from art why can't machines" or something to that effect is performatively conflating an organism and a machine.

My issue is with the rhetoric, if that isn't the rhetoric you are using I am not talking about you.

> these comparisons of llms with human artists copying are just ridiculous.

I've come to think of this as the "Performatively failing to recognize the difference between an organism and a machine" rhetorical device that people employ here and elsewhere.

The person making the argument is capable of distinguishing the two things, they just performatively choose not to do so.

[flagged]

I guess you were hoping that people wouldn't post opinions related to "democracy" in a discussion of a product offering "democratic AI rails."

You worked very hard to offer an "opinion" that largely ignored the topic of discussion.

The problem is your "opinion" in avoiding the topic of discussion says basically nothing.

Sci-Fi Author: Inspired by human atrocities, I present to you my new novel: God Emperor of Dune.

Tech adjacent blogger: Hey guys here me out I love that we're building "starships" but it would also be spiffy if we end democracy and appoint a God Emperor!

Nintendo did something like this with "Pacman vs" for GameCube. The multiplayer gimmick was the players controlling the ghosts couldn't see the entire map and the player controlling Pacman could with a Game Boy Advanced linked to the Gamecube.

Yes my suggestion is that negations can work just fine, depending on the model and task, and instead of avoiding negations you can try other promoting strategies like emphasizing what you want at the beginning and at the end of the prompt.

If you think negations never work tell Gemini 2.5 to "write 10 sentences that do not include the word the" and see what happens.

This is sort of LLM specific. For some tasks you might try including the word comment but give the order at the beginning and end of the prompt. This is very model dependent. Like:

Refractor this. Do not write any comments.

<code to refractor>

As a reminder your task is to refractor the above code and do not write any comments.

Digital Foundry said these cards are estimated to cost game companies 16 bucks per unit.

There was actually a similar thing on this website where someone here claimed in an article comment section that if you feed your wife's emails into ChatGPT it can simulate her well enough to tell you whether she is cheating.

The weird thing is I looked at the posting history of that commenter and it was bland tech comments with no obvious signs of mental health issues.

Except it's wrong about this being a classic riddle.

I'm not familiar with this blog but the proposed "universal jailbreak" is fairly similar to jailbreaks the author could have found on places like reddit or 4chan.

I have a feeling the author is full of hot air and this was neither novel or universal.

Unless I'm missing something you choose from a list of principal business Codes and also provide a description.

Unless "drug dealer" is one of those codes it's not an option to select that so that isn't the correct answer for code.

The instructions for the codes state, "Note that most codes describe more than one type of activity."

If you must provide a description as well you would provide one that describes more than 1 type of activity, not "drug dealer".

Merchant? Salesman? Deliveries? Distributor? Client services? Sales?

That field isn't actually used for anything other than determining if you are eligible for tax breaks like the school teacher deductable.

I'm not sure if this is intended to be a hypothetical but there is no IRS form where "Drug Dealer" is the correct answer.

Drug dealing income would be disclosed as "Other income".

If you volunteer "drug dealer" my guess is they could use it against you. Similar to showing up at FBI headquarters and shouting "I'm a drug dealer!"

Gemini 2.5 is not generating that refusal. It's a separate censorship model.

It's more clear when you try via AI studio where that have censorship level toggles.

Asimov was not savy at computers and found it difficult to learn to use a word processor.

Does "Global app installs" the term used at the link, mean installs by unique human beings the way you are using it?

It's unlikely the model knows its actual cutoff date. Try asking 2024 news- for example in my test it knows the January 2024 Oscar nominees.

On AI studio the model told me today is June 13 2024.

As someone who gets little to no benefit from human therapy this genre of "please talk to a therapist- it cures what ails ya!" internet post always leaves me with unanswered questions about where the poster is coming from.

With ray tracing Nvidia made a new technology that breaks everyone's PC performance then they sold the solution to that broken performance as a card upgrade with better ray tracing hardware. Even though the Ray tracing doesn't necessarily look any better, just different, we now have to upgrade cards to solve the broken PC performance Nvidia caused themselves...