nicce

It is unfortunate. I guess we also can blame the social pressure. Most want features and don't value quality or durability unless something breaks and is unusable. Competence is often measured with shipping speed, rather than service had zero bugs over its lifetime.

Nor offer to compensate for the damages...

Probably to save their skin if agent starts to do some unexpected things and bringing havoc. But I doubt that OpenAI models with normal subscription, for example, wont even work with this skill.

I would personally whitelist people with good enough reputation/contribution history. Might get the best from both worlds. Avoiding AI spam while saving time for the core contributors. If they just have capability to use it responsibly.

I don't know if that holds.

Neither OpenAI or Apple "owns" the Linux project. As long as there is enough evidence that code is real contribution and not stolen property (if you can even prove it), there is no reason to deny it.

They already get downsides if company gets fines or even goes to bankruptcy. You never know whether they invested based on information that was not true at all. Which is unfortunately too common.

Including investors is a bit much unless they encouraged or mandated some decisons that enabled this.

> Dario said a "persistent botnet" and even links to the wiki page for botnet.

> By definition that needs a command and control server, the ability to execute tasks on demand and regular pings to the C2.

I am not sure if you are agreeing or disagreeing, but I am saying just in case, that Claude can be remotely controlled from the cloud on any machine it is installed and set running, and can be commanded remotely.

Instead of using enormous time for trying to bypass safeguards with prompt injections, just use some open model and you would be done already. These regular OpenAI models are not usable to any real work that touches cybersecurity.

> As the pace is so fast, you change your mind’s state so frequently and quickly that it may lead to a (mild) burnout.

This is so true. Cognitive brute-force always when dealing with AI.

> I still keep a Windows machine around for gaming but I don't use it for anything more than that. All my serious work happens inside of MacOS.

I am happy to tell that gaming is in pretty good state in Linux these days.

That is not how it works, at least in a civilized country. The charges are not about agents, it is about operational responsibility and negligence in the company itself.

CEO is responsible for letting this to happen, not enforcing enough supervision, if not intentionally, then being grossly negligent. More severe if encouraging and letting this kind of agent research and operations happen at scale, while knowing that it can damage other systems and businesses.

Not anymore. Qwen 3.8 27B is so great that 5k investment gives you excellent speed and quality.

That would be really cool, if you can route your local model for it.

How is Sam Altman still on loose then

> Agree! My only concern is - is the judicial system fast enough, and resilient enough?

We already have the laws. It is just software. But somehow people are confused that it is not.

Having a philosophy is okay, but people must understand the social coencequences as well. Maybe similar example would be that people allow only assembly in their projects and compiled languages are forbidden. The impact is here likely similar on projects that are needed for something, and not just for fun.

> You're also wrong that AI isn't the problem: it is. AI is the problem because it's too addictive and easy to use, and we're not rational creatures but instinctive.

This is kinda moving the responsibility to the wrong place, and similar issue is on other topics. Is it the fault of the store if they put expensive things on areas where they are easy to steal? And then the store complains about lost property. Should we ban placing the things on poor places or criminalize thiefs and disallow thiefs to enter the store again?

I feel like people are just taking the easy shortcut when making these policies and banning AI completely. With AI, it is very easy to make sloppy code where nobody didn't put much thought on it. But AI can be really useful if you have the patience to use it incrementally and not by just asking it to one shot something.

Since on average the people don't have patience and don't act with responsibility, spaghetti gets everywhere. It may sound tempting to ban the AI-generated code, but the issue is not the AI, but the lack of responsibility and potentially the the lack of experience of the developers whom are contributing with this spaghetti slop.

The judgment should go for the end result and the author is responsible if the code is bad. Add exponential back-off to contributions if they are bad.

It is also possible to brute-force maintainers with too many things to review, but you can also limit the lines of code they can contribute, unless they have some sort of reputation.

Banning the AI just means that everyone is using it, but nobody is saying it aloud. Since it saves so much time when used correctly.

In many way it is, because it allows defeting many incorret arguments people in power make, or simply asking where is the cheapest product ”X” with one question. Or simply be more aware or any kind of scam people or companies try to do for you. It is pushing equal intellectuality quite hard with low effort.

Many claims but no clear evidence that they actually find significantly more severe issues compared to open models.

> Altman and Bubeck claim that because the proof used different scaling parameters and analytical steps, it's not related,

Is Altman even mathematican these days?

How the same people can have access to this information and be acknowledged? Completely different place.

They are completely different things with different consequences. Someone can point to the sky and prove that the sun is there.

I think that in the end it does not matter, anymore. It is inevitable that most of the text now and in the future is at least reviewed by LLM.

We should criticize whether text is poorly written, inaccurate, wastes words to get to the point and anything like. Saying that it is "written by LLM" just bypasses this and is not useful and misses the point. AI written text can be really good if used correctly. Criticize the content, don't speculate how it was written. If AI helps us to write better text, that is great. But often it is not good.

It is better blame the author for the bad text, so they get consequences and might do better next time, if the text is bad.

By the way, I think this piece of text was quite good.

All these "is this AI" checkers should be banned, as they never can reach 100% accuracy.

To have a really good harness, it seems that you need a separate one for every different project you are working with.

There is a saying that ideas are cheap. Only delivery matters.

The real advancement, not brute forcing mathematical equations like some company does and claims revolutionary breakthrough. How many ideas like this has come from AI companies' models?

I have been using it today the whole day and it is definitely better than Luna. Great that bench agrees.

1. Not storing secrets properly or using hardcoded secrets

2. Wild use of webviews/iframes sometimes easily propagates as XSS in phones

3. Incorrect client-side OAuth 2.0 configuration e.g. with schema-based redirect URLs.

4. Not supporting high-enough API versions, which may prevent some OS-related weaknesses

5. The list is actually very long. Just few top of my mind.