You can read it if you add archive.ph/ in front of the link (replace the www with it). It really is a well written article. I also don’t like paywalls, but I like good journalism and as they are still figuring out a way to earn money on the web I also understand their choice for these types of strategies.
There is a difference between telling and putting videos on the internet of you masturbating.
As someone who lives in Europe (Belgium), this sounds absolutely insane. If a prison were to do something like this here it would be all over the news! Simply disgusting.
While I do agree with a lot of the points made in the article, most of them seem rather basic stuff that everybody has been talking about since AI has started to rise. The tool isn’t necessarily the problem, it’s the capitalist incentives for which it is created and used that drive us further towards dystopia. I found it rather funny to read a former exec criticizing capitalism. I wonder what he’s up to now?
Also, both the year 2027 and the timespan of 15 years seem confidently pulled out of his backside :).
This is such a measured response, I love you for it!
This is quite an old video, and lately he’s been working on his communication issues. He fully admits that he’s been a dick many times. This all escalated a couple of years ago where he even took a few months off and focussed solely on getting this fixed and getting help I believe.
Not to say he’s perfect now, but I don’t think we’ll ever get this sort of catharsis from Elon, Trump, Bezos…
These books were purchased by them before being destroyed in the scanning process. I fail to see the issue with this specific case. Lots of artists buy stuff and irreversibly modify it. Are we going to be angry now at people who glue their puzzles or use parts of books for scrapbooking? If these were unique works there would be an issue, but I don’t think that truly unique pieces would be in their target group, as the destructive scanning is all about cost cutting and unique works cost a lot of money that they wouldn’t just destroy.
The fact that they use it for model training and later sell access to that model’s work is the shady part that has a severe whiff of plagiarism to it.
Isn’t this site proof that the internet doesn’t have to die as a whole? I mean, I agree with your sentiment, but I feel that this will mostly hit the middle to high traffic sites. The community based ones with organic discovery will remain OK I think. This might even evolve back to an early internet age of smaller hobbyist sites because there no longer is big money in the internet (apart from what google captures as “their” internet)
If I were to ask my Magic 8 Ball “Is the word ‘difinitely’ misspelled?” 100 times, it’s going to reply in the affirmative over 16% of the time.
This comparison makes no sense. Your example has a binary question. In that case, any system that replies correctly at even a rate of around 50% would be useless. However, the problem space in this scenario is way larger than 2 options and still way larger than 100 options. Being correct in even a small number of 100 attempts is still statistically significant.
The fact that an LLM is unable to reason and that it is based on statistics doesn’t change anything about this behavior. At the end of the day you get a tool that is able to point you to actual new information that you by yourself did not arrive at.
Imagine that you put a lot of effort in a better model specifically for vulnerability research and you get it up to a correctness rate of a mere 10%. I would gladly hire some programmers to sift through these reports and possibly find overlooked vulnerabilities.
This is literally the very first experiment in this use case, done by a single person on a model that wasn’t specifically designed for this. The fact that it is able to formulate a correct response at all in this situation impresses me.
It would be easy to criticize this if it were the endpoint and this was being advertised as a tool for vulnerability research, but as discussed at the end of the post, this “quick little test” shows both initial promising results and had the fortunate byproduct of actually revealing a new vulnerability. By no means is it implied that it is now ready for use in this field.
The issue with hallucinations is one that in my opinion is never going to be totally fixed. That is why I hate the use of AI as a final arbiter of truth, which is sadly how a lot of people use it (I’ll quickly ask ChatGPT) and companies advertise it. What it is good at however, is coming up with plausible ideas, and in this case having an indication for things to check in code can be a great tool to discover new stuff, as is literally the case for this security researcher finding a new vulnerability after auditing the module themselves.
I hate AI. Why?
However
I also took the time to read the original blog post, and it is a fascinating story.
The author starts out with using an existing vulnerability as a benchmark for ChatGPT testing. They describe how they took the code specific to the vulnerability and packaged it for ChatGPT, how they formatted the query and what their results were. In 100 runs only 8 correctly identify the targeted vulnerability, the rest are false positives or claim that there are no vulnerabilities in the given code.
Then they take their test a step further and increase the amount of code shared with ChatGPT so that it also includes stuff of the module that had nothing to do with the original vulnerability. As expected, this larger input decreases performance and also reduces the vulnerability detection rate for the targeted vulnerability. However, in those 100 runs, another vulnerability was described that wasn’t a false positive. An actual new vulnerability that the author didn’t know about was discovered. Again, the signal to noise ratio is very low, and one has to sift through a lot of wrong reports to get a realistic one, but this proved that it could be used as a useful tool for helping to detect vulnerabilities.
I highly recommend reading the blog post.
As much as I like to be critical about AI, it doesn’t help if we put our heads in the sand and act as if it never does something cool.
One of the very few good pee jokes, congrats!
That’s the stupid part. Serious criminals will always find a way to encrypt their data. Any communication channel can be used for encrypted communication if you do the encryption part yourself (like with an agreed upon passphrase or something similar). The only people who are hit by this are the average users.
This makes me so mad.
Isn’t that worse though? They use your computing power but they still get the data :). At least with cloud apps you get the benefit of not having to run the app locally (this was the idea behind low powered chromebooks).
How come you have such strong feelings against Firefox?
Not trying to start a fight, just an honest question. Choosing to rather stop using the internet all together seems extreme for a browser that I use daily, but maybe I’m missing something in your use case.
Crunchyroll, can weebs not stay in their bedrooms privately??
Also, 2dehands, which is the biggest second hand buying / selling site and app of Belgium. Quite a big one!
I am am continuing to extend the life of my 6 years old laptop for as long as possible, hopefully I will be able to buy a framework when it does eventually die on me. (It’s a semi-shitty clevo model but I recently replaced the battery and it’s kind of decent again)
While technically correct, they do have it in China itself, it’s a modified version called Douyin. It is more restricted, censored and tightly controlled.
I agree that it is a cyberweapon, but don’t think that it’s only used against foreigners, they use it just as much to observe and influence their own population.
Finally, I would like to point out that to a lesser extent this is also the case for a lot of USA owned social media and tech companies. Edward Snowden’s revelations for example indicate this. While the extent of government control and influence is much larger in China, I wouldn’t underestimate the influences of Meta, Google and Microsoft for example.
Xonotic is still quite active!
I am a huge proponent of privacy, anonymity and freedom of information and I also totally support this idea.
A public place where people have to speak as their public selves is useful to have and doesn’t mean other spaces have to be removed. Especially for official communication from public entities, I prefer them to use this than some other sites that are rampant with bots, imposters and the like.