I have good results with this prompt after every larger change: Now do a final code check. Is everything tidy and do the components adhere to the principle of separations-of-concerns. Is everything in an understandable and maintainable state? Do we make any assumptions that may not be true anymore? Is any code left over from previous edits or experiments that does not belong into the codebase? Is the documentation still representing the current state of code?
Sound like snake oil. How would this work? The app that the agent is developing needs access to the file, so access to it cannot be blocked. Just because read_file can not access it (I think current harnesses prevent reading .env files already), does not mean the contents will never be seen by the model.
Sometimes I have weird fantasies about a post apocalyptic world where factories burned down and people have to live with the tech that’s available. No network, just off site solar power or generators, only local devices. I think it’s interesting to think about how far we could get with this.
Does anyone have recommendations for novels, movies or video games with that topic?
This is not about information but about capital.
Even if we had free access to the weights of the best models in the world: who would be able to run them?
If that’s true, then it will be even cheaper to provide them as a subscription. Following your logic, every company would be running their own data centers instead of using cloud providers.
This will not happen. None of the existing apps people use daily on their phones have any incentive to support this. Social media wants the people to doomscroll, shopping apps and booking sites want to use their own dark patterns to make people believe they get a special discount if they buy _now_ and everything else just wants users to see the ads. Why on earth would they offer convenient hooks for AI chatbots?
You're absolutely right! "Why Does This Matter?" was a dead giveaway that this article was not written by a human — it was written by a large language model.
But really: "Why does this matter?" When looking at an article like this, I rarely read the text. This is just fluff no matter if AI-generated or hand written. The info is "there's a LEGO set of that ASML machine" and the picture of that set. That's all I want to know before clicking the back button.
But there is interesting information in lots of the comments. Like the quote from Ternus about Apple Maps in one of the comments. This gives relevant insight of how he thinks and how he might handle problems when he takes over.
Yes, but none of this applies to the local codex agent that runs when I tell it to and has access to my computer. Like: „scan this folder of PDFs and create an excel file with all expenses. Then enter them into my tax software.“ This needs access to very sensitive data and involves a quite complex handling of data. But the only attack vector I see is someone injecting prompts into my invoice files.
How many of these threat vectors are just theoretical? Don’t use skills from random sources (just like don’t execute files from unknown sources). Don’t paste from untrusted sites (don’t click links on untrusted sites). Maybe there are fake documentation sites that the agent will search and have a prompt injected - but I haven’t heard of a single case where that happened. For now, the benefits outweigh the risk so much that I am willing to take it - and I think I have an almost complete knowledge of all the attack vectors.
Services revenue is mostly just 30% from App Store Sales. This means every time a user clicks a pro account for ChatGPT or Claude on their phone, Apple makes more money than they could make with a self deployed model.
Nobody said that. But as you say, it's just a tool. Tools need to be used correctly. If tools are unintuitive, maybe that's due to the nature of the tool or due to a flaw in it's design. But either way, you as the user need to work around that if you want to get the maximum use out of the tool.
You're absolutely right! No, really: I've never had this problem of unprompted changes when I'm just asking, but I always (I think even in real-life conversations with real people) start with feedback: "Works great. What happens if..."
I think people having different styles of prompting LLMs leads to different model preferences. It's like you can work better with some colleagues while with others it does not really "click".
As a non native speaker, I sometimes use LLMs to search for a way to formulate my thoughts like I intend them to be received by the reader. I'd never just copy the verbatim LLM output somewhere, it always sounds blunt and not like me, but I gladly apply grammar corrections or better phrasing.
I'd normally not do this for a text of this length, but just for fun, here's what ChatGPT suggests:
As a non-native speaker, I sometimes use LLMs to help me find wording that conveys my thoughts the way I want them to be understood by the reader. I would never copy the output verbatim, because it often sounds blunt and unlike me, but I’m happy to use grammar corrections or improved phrasing.