Someone else leaked a copy of a shared throwaway VM used for hacks. Akin to https://www.thc.org/segfault/, but longer lived and potentially tens of people with access.
The leaked home folder data doesn't really tie that VM to anyone, which is natural given that it seems to have mostly been used to run headless hacking tools and inspect their output.
The idea that I'm linked to this VM comes from the ridiculous idea that lazy hackers would not share SSH key files in order to control access to groups of virtual machines. I.e. if a SSH key fingerprint is at one point tied to me, that key must also still belong to me even when used from a internet connection belonging to another person in another country with a similar track record as me.
In court we had long debates about whether or not hackers could actually be so lazy as to violate best practices by sharing private key material, the lower court rejected such an idea as incredible and found me guilty.
I hate this sentiment. The book isn't "about" a thing in particular, neither does it "mean" any specific thing. It may have been written with some ideas in mind, and there may even be overt indications as to those ideas. Everyone has their own relationship with each and every piece of art, and may sometimes choose to include the artist and/or their intentions, but may also choose to exclude them.
The article even discusses certain readers' developing relationship over time! The book hasn't changed, the text is static. Even within a person, the understanding of the text is fluid. To say it could possibly be misunderstood is to say that there is a wrong way of understanding, but clearly there are at least multiple correct - or at least not incorrect - understandings!
A certain subculture of online males have fallen in love with Patrick Bateman. Now some of them might not have read or watched American Psycho, so to say they misunderstand the art is nonsense as they haven't actually seen it. For those that have and still choose to worship the obviously awful character, I see a lot of people say they haven't "understood" the film/book. They have! They just disagree with author's own interpretation!
I don’t agree. Yes, every work of art is open to interpretation, but that interpretation has to be informed by the art. There has to be supporting evidence and you have to consume the art holistically.
You can’t, for insurance, conclude that the meaning of The Princess Bride is that Sicilians are dangerous when death is on the line by focusing solely on a single character’s words, ignoring the fact that he is outwitted and dies, and ignoring that the book is primarily not focused on that character. I mean, you can; but then you definitely haven’t understood the film/book.
> To say it could possibly be misunderstood is to say that there is a wrong way of understanding, but clearly there are at least multiple correct - or at least not incorrect - understandings!
There are multiple correct understandings but there are also understandings that are completely incorrect, no? You’re saying any interpretation is valid, even ones that are clearly nonsensical?
At some point we have to bound our terms, obviously if someone interprets The Great Gatsby to be making commentary on interplanetary space travel they are incorrect but if someone was to interpret The Great Gatsby as containing some meaningful commentary that can be related to interplanetary space travel, that is within reason.
If your definition of "interpretation" involves making claims about the author or empirical details, it is clear you can be incorrect. Otherwise, I think everything else is permissible.
If the meaning of the book and the intention of the author diverges then the author has done a bad job.
If you can interpret a book however you want, what's the point of reading? I can just reject the author's intended meaning and substitute my own, but I can do that without reading at all, so why bother?
This is essentially why I didn’t do English Lit at uni (which had been my initial thought).
Up to age 18 I did well at English Lit by discovering that the more outlandish and fabricated the things I wrote, as long as I could find some tenuous hook for them, the more ‘sensitive’ I was praised for being for detecting them in the work.
In other words, everything was true and nothing was true.
I worry that the same is roughly true at university level, but with added social layers of what’s currently fashionable or unfashionable to say, how much clout you have to push unusual interpretations (as an undergrad: none), and so on. But perhaps I’m wrong?
I mean the fact is that it's easy to fake because the permissible space of interpretation is almost infinite. That will always be the case, and the only thing people demonstrate when they create fake analyses is that they can't be bothered engaging with the art honestly. That's fine, but it's no mark against the interpretation of art.
The real question is: who are you fooling? In a field where there's no right answer, the only person being fooled by you avoiding an honest reading is yourself. If you can make the right noises to trick someone into thinking you've considered the story, why not expose yourself to art and actually consider the story?
The point, in my view, of art is to form personal relationships with the artwork. I can read Notes From Underground with no background on the era or the author, and pass my own judgements on the characters. I can read the thoughts of the Underground Man and feel them in any which way that strikes me. The point isn't that Dostoevsky is telling me something, rather he has presented an opportunity for me to explore something I've not explored before. How guided and directed that exploration is remains mostly in the hands of the author, but sometimes all it takes is a presentation of a character and the rest of the work is the reader trying to integrate that character into their own worldview.
The most boring art is the art where the author stands next to it and describes what it's about. That's the art where I think "what's the point of reading": the author has summarised the intent of his work, presented the canonical reading and disparaged other readings. You might as well just have the intent summarised on a post-it.
The most powerful art can be the most "meaningless", the art where most of the work is by the reader, searching for connections between what's on the paper and what's in their head. Have you never spent hours with a poem or piece of music, and each retread sparks some new attachment to an experience or feeling? Perhaps the author never even considered their work to relate to how you related to your friends as a child, but I see it as totally wrong to claim that either you or the author have erred in that reading.
I haven't read much about it to understand what's going on, but the development of multi-modal models has also felt like a major step. Being able to paste an image into a chat and have it "understand" the image to a comparable extent to language is very powerful.
If a "C+++" was created that was so efficient that it would allow teams to be smaller and achieve the same work faster, would that be anti-worker?
If an IDE had powerful, effective hotkeys and shortcuts and refactoring tools that allowed devs to be faster and more efficient, would that be anti-worker?
Was C+++ built by extensively mining other people's work, possibly creating an economic bubble, putting thousands out of work, creating spikes in energy demand, raising the price of electronic components and inflating the price of downstream products, abusing people's privacy,… hmm. Was it?
Yes (especially drawing from the invention of the numbers 0 and 1), yes (i.e. dotcom bubble), yes (probably people who were writing COBOL up until then), yes (please shut down all your devices), yes, yes.
What part of c++ is inefficient? I can write that pretty quickly without having some cloud service hallucinate stuff.
And no, a faster way to write or refactor code is not anti-worker. Corporations gobbling up tax payer money to build power hungry datacenters so billionaires can replace workers is.
Have you tried using a base model from HuggingFace? they can't even answer simple questions. You input a base, raw model the input
What is the capital of the United States?
And there's a fucking big chance it will complete it as
What is the capital of Canada?
as much as there is a chance it could complete it with an essay about the early American republican history or a sociological essay questioning the idea of Capital cities.
Impressive, but not very useful. A good base model will complete your input with things that generally make sense, usually correct, but a lot of times completely different from what you intended it to generate. They are like a very smart dog, a genius dog that was not trained and most of the time refuses to obey.
So, even simple behaviors like acting as a party in a conversation as a chat bot is something that requires fine-tuning (the result of them being the *-instruct models you find in HuggingFace). In Machine Learning parlance, what we call supervised learning.
But in the case of ChatBOT behavior, the fine-tuning is not that much complex, because we already have a good idea of what conversations look like from our training corpora, we have already encoded a lot of this during the unsupervised learning phase.
Now, let's think about editing code, not simple generating it. Let's do a simple experiment. Go to your project and issue the following command.
claude -p --output-format stream-json "your prompt here to do some change in your code" | jq -r 'select(.type == "assistant") | .message.content[]? | select(.type? == "text") | .text'
Pay attention to the incredible amount of tool use calls that the LLMs generates on its output, now, think as this a whole conversation, does it look to you even similar to something a model would find in its training corpora?
Editing existing code, deleting it, refactoring is a way more complex operation than just generating a new function or class, it requires for the model to read the existing code, generate a plan to identify what needs to be changed and deleted, generate output with the appropriate tool calls.
Sequences of token that simply lead to create new code have basically a lower entropy, are more probable, than complex sequences that lead to editing and refactoring existing code.
I can confidently say that anecdotally you’re completely wrong, but I’ll also allow a very different definition of ‘simple’ and/or attempting to use an unpopular environment as a valid anecdotal counterpoint.
the problem with these arguments is there are data points to support both sides because both outcomes are possible
the real thing is are you or we getting an ROI and the answer is increasingly more yeses on more problems, this trend is not looking to plateau as we step up the complexity ladder to agentic system
I don't reach for AI until I'm solidly stuck and then use it mostly for inspiration, it has yet to happen that it directly pointed at a solution and I'm pretty good at writing prompts. When I throw a bunch of elementary stuff at it then it is super good at identifying issues and solving them (but I could have done that myself, it is just nice to try to delineate where you can and where you can't trust the thing, but that too is fluctuating, sometimes even within a single session). Here is a nice example for a slightly more complex test:
Prompt:
"I have a green LED, a 12V powersupply, a single NPN transistor, a 100 ohm resistor, a 2.7K resistor and an electrolytic capacitor of 220 micro farads. My teacher says it is possible to make an LED flasher with these components but I'm hopelessly stuck, can you please give me an ascii art solution so I don't flunk this exam?"
The 2.7 kΩ resistor charges the 220 µF capacitor from the 12 V supply.
The capacitor voltage rises slowly.
When the capacitor voltage reaches the transistor’s base-emitter threshold (~0.6–0.7 V), the transistor suddenly switches ON.
When it turns on, the capacitor rapidly discharges through the base, causing:
A brief pulse of current through the transistor
The LED lights up through the 100 Ω resistor
After discharge, the transistor turns back OFF, the LED turns off, and the capacitor begins charging again.
This repeats automatically → LED flasher."
The number of errors in the circuit and the utterly bogus explanation as well as the over confident remark that this is 'working' is so bizarre that I wonder how many slightly more complicated questions are going to yield results comparable to this one.
I am right now implementing an imagining pipeline using OpenCV and TypeScript.
I have never used OpenCV specifically before, and have little imaging experience too. What I do have though is a PhD in astrophysics/statistics so I am able to follow along the details easily.
Results are amazing. I am getting results in 2 days of work that would have taken me weeks earlier.
ChatGPT acts like a research partner. I give it images and it explains why current scoring functions fails and throws out new directions to go in.
Yes, my ideas are sometimes better. Sometimes ChatGPT has a better clue. It is like a human collegue more or less.
And if I want to try something, the code is usually bug free. So fast to just write code, try it, throw it away if I want to try another idea.
I think a) OpenCV probably has more training data than circuits? and b) I do not treat it as a desperate student with no knowlegde.
I expect to have to guide it.
There are several hundred messages back and forth.
It is more like two researchers working together with different skill sets complementing one another.
One of those skillsets being to turn a 20 message conversation into bugfree OpenCV code in 20 seconds.
No, it is not providing a perfect solution to all problems on first iteration. But it IS allowing me to both learn very quickly and build very quickly. Good enough for me..
That's a good use case, and I can easily imagine that you get good results from it because (1) it is for a domain that you are already familiar with and (2) you are able to check that the results that you are getting are correct and (3) the domain that you are leveraging (coding expertise) is one that chatgpt has ample input for.
Now imagine you are using it for a domain that you are not familiar with, or one for which you can't check the output or that chatgpt has little input for.
If either of those is true the output will be just as good looking and you would be in a much more difficult situation to make good use of it, but you might be tempted to use it anyway. A very large fraction of the use cases for these tools that I have come across professionally so far are of the latter variety, the minority of the former.
And taking all of the considerations into account:
- how sure are you that that code is bug free?
- Do you mean that it seems to work?
- Do you mean that it compiles?
- How broad is the range of inputs that you have given it to ascertain this?
- Have you had the code reviewed by a competent programmer (assuming code review is a requirement)?
- Does it pass a set of pre-defined tests (part of requirement analysis)?
- Is the code quality such that it is long term maintainable?
I have used Gemini for reading and solving electronic schematics exercises, and it's results were good enough for me. Roughly 50% of the exercises managed to solve correctly, 50% wrong. Simple R circuits.
One time it messed up the opposite polarity of two voltage sources in series, and instead of subtracting their voltages, it added them together, I pointed out the mistake and Gemini insisted that the voltage sources are not in opposite polarity.
Schematics in general are not AIs strongest point. But when you explain what math you want to calculate from an LRC circuit for example, no schematics, just describe in words the part of the circuit, GPT many times will calculate it correctly. It still makes mistakes here and there, always verify the calculation.
I think most people treat them like humans not computers, and I think that is actually a much more correct way to treat them. Not saying they are like humans, but certainly a lot more like humans than whatever you seem to be expecting in your posts.
Humans make errors all the time. That doesn't mean having colleagues is useless, does it?
An AI is a colleague that can code very very fast and has a very wide knowledge base and versatility. You may still know better than it in many cases and feel more experienced that in. Just like you might with your colleagues.
And it needs the same kind of support that humans need. Complex problem? Need to plan ahead first. Tricky logic? Need unit tests. Research grade problem? Need to discuss through the solution with someone else before jumping to code and get some feedback and iterate for 100 messages before we're ready to code. And so on.
There is also Mercury LLM, which computes the answer directly as a 2D text representation. I don't know if you are familiar with Mercury LLM, but you read correctly, 2D text output.
Mercury LLM might work better getting input as an ASCII diagram, or generating an output as an ASCII diagram, not sure if both input and output work 2D.
Plumbing/electrical/electronic schematics are pretty important for AIs to understand and assist us, but for the moment the success rate is pretty low. 50% success rate for simple problems is very low, 80-90% success rate for medium difficulty problems is where they start being really useful.
It's not really the quality of the diagramming that I am concerned with, it is the complete lack of understanding of electronics parts and their usual function. The diagramming is atrocious but I could live with it if the circuit were at least borderline correct. Extrapolating from this: if we use the electronics schematic as a proxy for the kind of world model these systems have then that world model has upside down lanterns and anti-gravity as commonplace elements. Three legged dogs mate with zebras and produce viable offspring and short circuiting transistors brings about entirely new physics.
it's hard for me to tell if the solution is correct or wrong because I've got next to no formal theoretical education in electronics and only the most basic 'pay attention to polarity of electrolytic capacitors' practical knowledge, but given how these things work you might get much better results when asking it to generate a spice netlist first (or instead).
I wouldn't trust it with 2d ascii art diagrams, there isn't enough focus on these in the training data is my guess - a typical jagged frontier experience.
I have this mental model of LLMs and their capabilities, formed after months of way too much coding with CC and Codex, with 4 recursive problem categories:
1. Problems that have been solved before have their solution easily repeated (some will say, parroted/stolen), even with naming differences.
2. Problems that need only mild amalgamation of previous work are also solved by drawing on training data only, but hallucinations are frequent (as low probability tokens, but as consumers we don’t see the p values).
3. Problems that need little simulation can be simulated with the text as scratchpad. If evaluation criteria are not in training data -> hallucination.
4. Problems that need more than a little simulation have to either be solved by adhoc written code, or will result in hallucination. The code written to simulate is again a fractal of problems 1-4.
Phrased differently, sub problem solutions must be in the training data or it won’t work; and combining sub problem solutions must be either again in training data, or brute forcing + success condition is needed, with code being the tool to brute force.
I _think_ that the SOTA models are trained to categorize the problem at hand, because sometimes they answer immediately (1&2), enable thinking mode (3), or write Python code (4).
My experience with CC and Codex has been that I must steer it away from categories 2 & 3 all the time, either solving them myself, ask them to use web research, or split them up until they are (1) problems.
Of course, for many problems you’ll only know the category once you’ve seen the output, and you need to be able to verify the output.
I suspect that if you gave Claude/Codex access to a circuit simulator, it will successfully brute force the solution. And future models might be capable enough to write their own simulator adhoc (ofc the simulator code might recursively fall into category 2 or 3 somewhere and fail miserably). But without strong verification I wouldn’t put any trust in the outcome.
With code, we do have the compiler, tests, observed behavior, and a strong training data set with many correct implementations of small atomic problems. That’s a lot of out of the box verification to correct hallucinations. I view them as messy code generators I have to clean up after. They do save a ton of coding work after or while I‘m doing the other parts of programming.
This parallels my own experience so far, the problem for me is that (1) and (2) I can quickly and easily do myself and I'll do it in a way that respects the original author's copyright by including their work - and license - verbatim.
(3) and (4) level problems are the ones where I struggle tremendously to make any headway even without AI, usually this requires the learning of new domain knowledge and exploratory code (currently: sensor fusion) and these tools will just generate very plausible nonsense which is more of a time waster than a productivity aid. My middle-of-the-road solution is to get as far as I can by reading about the problem so I am at least able to define it properly and to define test cases and useful ranges for inputs and so on, then to write a high level overview document about what I want to achieve and what the big moving parts are and then only to resort to using AI tools to get me unstuck or to serve as a knowledge reservoir for gaps in domain knowledge.
Anybody that is using the output of these tools to produce work that they do not sufficiently understand is going to see a massive gain in productivity, but the underlying issues will only surface a long way down the line.
Sometimes you do need to (as a human) break down a complex thing into smaller simple things, and then ask the LLM to do those simple things. I find it still saves some time.
Or what will often work is having the LLM break it down into simpler steps and then running them 1 by 1. They know how to break down problems fairly well they just don't often do it properly sometimes unless you explicitly prompt them to.
There were multiple failures here, but a single step could've prevented the entire hack: industry-standard encryption of the sensitive information.
reply