cmiles74

It seems 50/50 to me that an LLM would come up with a similar solution.

Or things will settle down enough that people start looking at the numbers and doing the math. LLM tooling can generate code and often that code is correct. In my experience it takes about as long for the tooling to generate the code as it would for me to write the code myself and it also costs more money (my time plus the token cost). The more complicated the task the more likely I have to go back and get the model to re-work the majority of it as it made some assumptions that, after a little thought, turn out to be problematic.

IMHO, at the end of the day it's been about the same amount of time, the cost is a little higher and I don't feel as confident in the end product. I suspect that I am not the only one that feels this way!

Vibe coded or not I think human hubris and incompetence played a big role here.

I was thinking that this is exactly what ParticleDB is: a random amateur vibe coding a DB. I wouldn't count on things changing much in a couple of years, people will still be people.

I'm on my phone, "If that’s hostile works.." should be "If that’s how it works..."

Cursor should not be shipping extensions that are branded Anthropic if Anthropic didn’t create that extension. If that’s really what happened then that needs to change, ASAP. Now that it has happened, IMHO, it is partly an Anthropic problem: the malware has their branding and their name on it, I don’t think they can just brush that off.

Does Cursor have some process that just scrapes Microsoft’s extension catalog and lists them in their own? If that’s hostile works then it’s broken and needs to be fixed.

I remember this bit from "Silicon Valley".

https://techcrunch.com/2017/02/19/when-less-is-more/

Bad prompting plus a weak review process?

I wasn't wow-ed by this piece either. There does seem to be something there, though: this thing where someone commits some serious fraud yet gives Trump a slice and then they're pardoned, along with the precipitous drop in prosecuting just these kinds of cases.

Maybe this Atlantic story? https://www.theatlantic.com/magazine/archive/2019/05/the-tro...

I suspect this is more marketing than fact. I don't think there's any real evidence that this is the case, just hints dropped by OpenAI and Anthropic.

I believe there are software developers out there that might take a stronger position against AI but they literally fear for their employment.

I think this is a stretch, many engineering feats are also been considered works of art. Bridges, homes, office buildings, etc. There is no reason not to consider software code, or the software itself, to be works of art.

Indeed, we have seen many instances of software that are well regarded as being works of art. Is the source code for that software somehow not art?

I'm sure that is true. On the other hand, the majority of their customers are likely using these tools for work and larger organizations are paying for the licenses. People in school for the fields that use these tools are probably being trained on Adobe products, with some discount for the school and the students. I don't think these vibe-coded apps are going to get much traction, there's a whole entrenched ecosystem around Adobe products.

My distaste for vibe coded apps aside, I wish them the best of luck.

Reading it over, it does seem like giving the app full disk access would be enough for it to read our messages. I mean, they are stored on disk somewhere.

I gotta' disagree on this one. Meta made claims that it was taking privacy seriously and it turns out, not so much. I do think they should be getting some pressure on that score.

They gave iSight cameras away at the conference. They were amazing, I used mine for years.

Maybe OpenAI is serious about securing the environment they run their models in, but then again maybe not. I don’t think we can tell from here.

IMHO, I haven’t been super impressed with the security measures I’ve had to work with. Often they are simplistic and bolted on at the very end. If it comes to light that this is the attitude OpenAI has been taking, I would not be surprised.

An LLM, in my opinion, is not comparable to a person.

On the risk management angle, for sure it’s a spectrum. I don’t agree that the far end of the safe side of that spectrum for AI models is “entirely safe and entirely useless”, there is a lot of work you can do with a model that has zero risk of hurting anyone (aside from your wallet). If someone chooses a more dangerous spot on that spectrum, I believe they should be held responsible.

> An AI given the same exact instructions and tools can go and complete a task you wanted it to. Or it can get sidetracked into breaking out of your sandbox and hacking Pentagon. No way to know in advance.

This has not been my experience. I’ve been getting a lot of good work done and, as of today, have been involved in zero Pentagon hacking incidents. ;-)

An argument can be made that “you never know” how the AI model might respond to something. OTOH, someone has to decide what tools to give it; maybe don’t provide dangerous tools to an unpredictable LLM.

This all seems like a way to try to avoid taking responsibility for the model’s actions. Someone puts the tools in place, someone provides the instruction and, sometimes, someone decides not to monitor the model’s output.

I’m not sure what I’m missing, isn’t it what you hook the LLM up to and the instructions a person gives the model that makes it dangerous? Claiming this is an inherent quality of the tool itself seems kind of off-the-rails to me.

IMHO, if the model breaks a law, apply the law to the operator.

Maybe they think investors will accept this and let these fees slide in the hopes that the AI bubble can be milked for a little longer.

Could this be a sign of the end of the AI bubble?

What would truly be user hostile would be to allow so many processes per UID that a small handful (maybe using VSCode) make the system slow or unresponsive to everyone else.

In my opinion, this article is talking about people who get into long conversations and become convinced the model is “intelligent”, maybe even “AGI”. This seems to be a trap people are falling into, even in 2026.

Anthropomorphizing the tool led directly to this problem, where we nearly started a war with China.

Or maybe it should be a flag to indicate that there might be more to think about before handing the task of to an LLM.

Skilled and dedicated people and an organization dedicated to maintaining a consistent level of security. Maintaining both of these long term seems to be the challenge for many companies.

All the better, I've long suspected that these companies are collecting this data and selling some portion of it. Having this category of business entirely disappear sounds like a solid win to me.

Will the LLM tools be able to maintain this pile of code either? In my experience there seems to be some complexity maximum to a project that, once reached, the LLM tooling doesn't seem able to manage it either.