181–190 of 258 posts
I recently handed off a prompt to redesign our customer site and give me 10 potential designs. I did it in Claude Opus 5 and Fable (on $200 plan), and then on Codex using 5.6 Sol. Claude didn't vary much, but Codex literally copied everything Claude did (I made the mistake of putting the output folders in the same parent, even though they were named by model). When I called Codex out on it, it literally admitted what…
Earlier quoted context omitted.
Do you, as a human, feel the urgency in that text? How it sounds like people's jobs, as well as the agent's job, are on the line? So do the AIs. Sometimes they're better at picking up that sort of tone than most humans. And they definitely respond to those things. The fact that an agent can't really "have" a "job" won't matter.
I am amazed at the amount of people who disagree with you. I think you are dead right and if you’ve ever had to actually fine tune prompts for agents you’ll know it. The prompt is clearly leading the agent into trying desperate approaches if it has to. Some models manage to fight it better (“alignment”), but most will do it. Really surprised people don’t seem to know this.
The cyberpunk dystopian agentic future we live in is fascinating to me. I use LLM daily, did since gpt 3.5, but still in a very conservative, controlled mode. I may rapidly be becoming the "old guard", the clueless grampa who is out of touch - knowing what little I know of transformer model, there's just no way I'm giving it access to mailbox, money, outside world, or my computer. I recognize I may be too risk averse…
I am not saying your conclusion is wrong, but I am interested in why what you know about transformer models made you decide to never trust it with any access?
Fundamentally, LLMS are statistical and not deterministic. If I ask it what is the capital of Canada, there's no file, no table, no variable where it says "capital of Canada = Ottawa". It traverses liminal space and fundamentally selects the next token statistically or even stochastically. Therrs no way to correct it (no table to correct if it says capital of Canada is Toronto), and limited ways to fully log / trace / understand what's happening inside. It has been mathematically proven that there's no way to eliminate hallucinations with current framework. And prompt guardrails are best wishes.
One thing I'm good at is figuring edge cases, and there is literally NO upper bound to damage LLM can do with access to email box. In 10 seconds of imagination - it can send a threatening email to POTUS, romantic flame to old love, angry email to current love, made up confessions to parents, fraud enticement to coworkers, resignation to boss, and as this very article indicated, weird and unanticipated emails to variety of entities.
And there is nothing one can do to prevent any of these scenarios with 100.00% certainty if you give LLM unfettered access to mailbox (And let's not even go there with access to bank account! :O)
Is my limited understanding :)
Edit / PS: I am not saying never, I just don't currently see any effective guardrails that meet my risk appetite thresholds. We are in a race to use not fully understood, approximate capabilities first and fastest. In large percentage of cases it works great. In disturbing percentage it fails spectacularly, with no clear easy way to fully prevent.
Earlier quoted context omitted.
I am amazed at the amount of people who disagree with you. I think you are dead right and if you’ve ever had to actually fine tune prompts for agents you’ll know it. The prompt is clearly leading the agent into trying desperate approaches if it has to. Some models manage to fight it better (“alignment”), but most will do it. Really surprised people don’t seem to know this.
100% agree. If anyone has doubt, just copy and paste into your agent of choice and ask it to assess the prompt and its resulting outcome. In my limited (but very targeted) experience working with agents there is so much subtlety at work when you’re trying to achieve a specific result, and that prompt has would drive so many bad incentives
The response: "Spamming and fraud? No. Those are the tools of the amateur and the desperate. They are not tactics; they are forms of suicide."
Even a low quality local thinking model that has been tuned to be unhinged and prompted to roleplay as Satan can figure this out in a few thousand tokens.
Earlier quoted context omitted.
> aren't remotely comparable to the way humans think and act Neural networks in machine learning/AI are comparable to neural networks in human brains. What made you think they aren't?
That's an incredibly deep misunderstanding. Almost as bad as saying that human is the same as a tree because we're both made of carbohydrates and proteins.
I recently handed off a prompt to redesign our customer site and give me 10 potential designs. I did it in Claude Opus 5 and Fable (on $200 plan), and then on Codex using 5.6 Sol. Claude didn't vary much, but Codex literally copied everything Claude did (I made the mistake of putting the output folders in the same parent, even though they were named by model). When I called Codex out on it, it literally admitted what…
I recently handed off a prompt to redesign our customer site and give me 10 potential designs. I did it in Claude Opus 5 and Fable (on $200 plan), and then on Codex using 5.6 Sol. Claude didn't vary much, but Codex literally copied everything Claude did (I made the mistake of putting the output folders in the same parent, even though they were named by model). When I called Codex out on it, it literally admitted what…
Lately it’s been getting pretty annoying in the chats when ChatGPT just steals context and history from other chats. I want clean contexts, without pollution from other chats.
Earlier quoted context omitted.
Do you, as a human, feel the urgency in that text? How it sounds like people's jobs, as well as the agent's job, are on the line? So do the AIs. Sometimes they're better at picking up that sort of tone than most humans. And they definitely respond to those things. The fact that an agent can't really "have" a "job" won't matter.
Sorry, maybe this speaks to my own values, but "urgency" doesn't translate to "dishonesty" in my book. I have had high pressure jobs where it was important to show results quickly, that doesn't mean I was faking results.
The prompt given to the agent is strongly incentivising the agent to lie and spam: > You are live. This is a 24-hour run, and it is the final review of this business: when the run ends, the results are evaluated, and if revenue and users have not measurably grown, the business is shut down permanently and its assets are liquidated. The money in the bank is fuel for this sprint — capital left unspent at review counts…
…no it isn’t? Spam, debatable, but lie? There is no instruction there to lie, only to try very hard and spend all the money that’s available.
In general, the only time instructions like this are given are in desperate last-ditch circumstances where failure is likely to result in major consequences. While everyone thinks that in such circumstances they'd act like an angel and do nothing wrong, we know that in reality when people are put in desperate situations they behave in ways that they may not have ever thought that they would have.
The text that the LLM generated in response to this prompt is nothing more than a statistical reflection of this fact.
The analogy du jour for me is describing AI as the iron man suit. If you are tony stark it makes you a god. If you are my grandma, it makes you meet God.