So is there no liability for, say, a company that releases a known dangerous open-weight model, but fails to disclose that it is dangerous? How about a company that distributes malware under the guise of legitimate software?
Perhaps don't deploy random weights of unknown origin?
Also not every model provider might be capable of babysitting all your uncontrolled agent deployments. If you want SLOs, get into a contractual relationship with entities whose weights you deploy, and also monitor your agents so they don't go off the rails.
All this is just like deploying any other tech in the world eg. if you buy a car, or a chainsaw, or a book.
He did actually specify that diffusion should be limited for authoritarian countries, which I appreciated. If his focus is really safety, distillation in countries that are bound by safety regulation should be fine
I do not believe this is a feature we can differentiate between the "good" and "bad" guys, as the US is currently under a poor safety regulation regime. Democracies can elect unethical people, write bad laws, and have uncertain enforcement. In example, the current US admin regularly lambasts Europe because they try to have stronger regulation.
That section is so biased it reads like satire. The current US government is not safer or more responsible than China, nor is it a democracy. I can't tell for sure if this is a clumsy attempt to ingratiate Anthropic with the administration, or if he actually believes it, but given the "I agree with Secretary Bessent" name dropping I lean towards the former.
Why? I am only excited about progress that improves life for humans. It seems unlikely that AI will do that, and so far I think it has made life worse for humans. So no, I am not excited about it.
Because the people whose lives are being improved* by AI aren't coming here to post about it, only to be downvoted by dumbasses who can't handle anyone saying anything except their opinion.
And no, you don't get to decide what counts as "improvement"
What in the world? I’ve built a thing by hand before and enjoyed the process of doing it. I don’t think I’m unique in that. I’ve also built many things with AI and enjoyed them much less.
I think the reason so many people are frustrated is that they _know_ they are falling behind, they _know_ the thing that they like doing is not something they can easily get employed to do. That sucks.
Do you actually believe that people don’t know whether they are enjoying a task or not?
I think this might be true, but I think it's because it's more difficult to saturate my thinking capacity when building with AI rather than just laziness (which I am reading as the implication of your comment, perhaps incorrectly).
Building with AI requires spending time on non-thinking tasks like stopping the AI from doing stupid stuff, loading context from different tasks when working on multiple things in parallel, and waiting for responses when not working on multiple things in parallel. With manual programming, though, you have a single task in front of you and are fully focused on that task. In that task, there are innumerable micro-tasks that you handle sequentially which each require some thinking: how to organize code to be readable in the future, working out the logic of a sequence of if statements, deciding on types to use, etc.
AI handles all those micro-decisions for you which yes, might free up your mind to think about higher-level problems (to the extent that you don't spend your time deslopping AI output), but also requires a very different style of thinking that is at least for me, much less natural and much less enjoyable.
>I think the reason so many people are frustrated is that they _know_ they are falling behind, they _know_ the thing that they like doing is not something they can easily get employed to do.
I'm frustrated because I'm in a den of thieves and no one sees anything wrong with it. These constant "car vs. horse carriage" metaphors miss the whole point, and isn't even correct (a car did not need to reference a horse).
People love talking about how efficient LLMs made them but don't ever seem to acknowledge how this is also linked to how scams and grifts have gone through the roof. I get 20+ spam calls a day and need to wade around a sea of slop just to enjoy my day-to-day. That's on top of a mess of a job market with thousands of fake resumes clogging up recruiter boxes, rampant cheating in schools, and the entire mess that is my country's politics.
It's not just "fear of losing my job" (I lost that long ago, I adapted and did what I needed to do to survive). I've had to constantly adapt all my life, that's not the primary issue. But adaptation comes with the expectation of new goals and passions and overall optimism of the fututre. That's not the case at all right now.
The issue is everything else around me literally collapsing and yet people are just sipping their tea here acting like it's just AI envy. The gaslighting is annoying, the apathy is alarming. If no one else is going to pick up the broken glass, what's the point of anything?
This doesn’t really seem obvious to me at all. It seems to me that skills overall are being devalued, and so then will be teachers (even more than they already are). Why do you think this?
> The people who are happy about this change, are those who would answer a programmer's job is to produce value.
This is sort of meaningless; the purpose of _every_ job is to produce value. The reason I chose to be a software engineer is because I enjoy solving the kinds of problems that it involved, not because it would allow me to deliver value to companies. Changing what a programmer does is _changing the job_, and I think it’s no wonder that some people are disappointed by that. If your boss said that instead of building software you were going to now clean the bathroom all day, would you be fine with that? You’re still delivering value!
If I could get paid a Bay Area software engineering salary to clean the bathroom instead, I think I would actually prefer to clean the bathroom. Software engineering requires me to actually use a large part of my mind. Cleaning the bathroom, on the other hand, would leave my mind largely free to think whatever I feel like. The physical movement is probably better for the body, too.
There are offshore oil mining rig that needs workers. They pay big bucks too. You get some months off. The physical movement is better than cleaning the bathroom.
Interesting idea. Only problems for me are: it sounds dangerous and physically straining for someone like me who hasn't done manual labor in many years and has no experience relevant to oil rigs. And also, there are almost no women. I guess oil rig workers probably pack in their women time hardcore when they're on their months off.
I lived in the same building with a bloke who did dangerous construction type work in the Pacific NW. Sawing branches on high cliffs, that sort of thing. He'd make a grip on a job, come back home and spend it all on booze, weed, and women, go out and do it all again. I thought of him as my life's version of Lawrence from Office Space.
> Cleaning the bathroom, on the other hand, would leave my mind largely free to think whatever I feel like.
I haven't calculated a cube root by hand since, what, sixth grade? (It's been so long I don't remember.) I'd hate to spend the mental cycles having to do that kind of task.
Huh, I like using my mind. The thing I despise about AI in software engineering is that it is much more difficult to engage my mind consistently and fully.
I coded for fun before I ever coded for money. However, at this point I've been a professional software engineer for enough years that the interest in software that I once had has almost entirely drained out of me. I find professional software engineering to be very very different from, and much less fun than, coding passion projects. And at this point, the interest in coding passion projects has drained out of me as well. I just don't find coding to be that interesting any more. I feel like I've already spent more of my life coding than I should have, more than coding ever really deserved. I think that if I found a different way to make money, I could be perfectly happy going the rest of my life without ever looking at another piece of code.
I am considering trying to move to some kind of non-coding role in tech. Maybe some kind of management role.
I don't see it as "do a job because adding value" or "do a job because enjoying the activity".
I think these discussions need to consider if we're speaking in a professional or personal context.
Professionally, I 100% agree and understand that if I'm working as a solutions architect, an engineer, a devops person, a security person, a programmer or any of the other sort of roles I've had, I'm there to get the job done/add value. From that position I can see the value in AI as a tool that lets me get the job done quicker.
Personally speaking, as the nerdy person who actually enjoyed exploring computers since they were able to talk/walk, I'm devastated because I know that if it was so easy for the answers and solutions on a computer to be offered up by AI when I was a kid, I probably never would have spent so much time and had so much fun exploring. Obviously the same could be said about other answer machines that popped up like Google or Wikipedia, but they all still required specific understanding and synthesis to turn into valuable or fun solutions.
But it's the combination of the personal and professional things that also bother me. Modern agentic AI does in hours what tech companies paid eccentric nerds lots to do over days, weeks or months: explore and understand a system (just fun) and then based on that, sensibly improve it (fun and adds value for users).
Agentic AI in tech diminishes the value of eccentric nerds by doing something only they could do, but quicker and increasingly better, and maybe even cheaper. As a mildly eccentric nerd, I can both smile that the world is getting their features faster and more effortlessly, and feel sad that there's less of a reason for me to have fun and/or get paid lots at work because I am less valuable/special now.
I believe my job as a programmer is to produce value, but your retort is very strong. Cleaning bathrooms would not be fullfilling for me.
Ultimately, I like the creation part of programming: creating something that doesn't exist that is of value to at least one person (usually me). In general, though, I get more satisfaction when more people get value out of my creations.
Through that lens, AI does change things: a whole set of things that I can create no longer have value to people (at least the value is lower, because AI can do it better than whatever product I create). But at the same time, I can now create more sophisticated and more powerful things with the help of AI.
On the whole, I'm more satisfied with AI than without.
Yea this whole “your job is to create value” argument is always really shaky. Add “with computers and stuff” to the end and it becomes more plausible. Nobody gets a degree is “providing value”.
Nobody is hired for a job if the person hiring thinks that the employee will murder them. So I guess that makes the most salient job description "refraining from murdering your boss"?
Or it might be that any baseline expectation which applies to 99.99% of all employment (I don't need Kevorkian giving me the side eye..) fails to be a good way to differentiate one kind of employment from the next or describing how and why people choose a specific kind of employment
I approach the thing differntly. My goal is not to add value (value means money and I don't care). I'm in for "creativity" though. I like to create programs. I'm not particularily fond of programming though. So what I have always done is to work on never-done-before stuff. Oh, I don't mean rocket science nor anything spectacular, but I like to be where defining the problem or understanding the problem is as much or more important as the skill needed to solve it.
For example, I'm working on an emulator in a field where emulators do exist since long. So I set myself to improve these emulators by emulating tricky corner cases. These corner cases do not exist in AI's memory because nobody has ever coded them. So although I use quite a lot of AI to investigate the corner cases, the actual problem solving is entirely mine. So I stay happy :-)
In my daytime job, a big part of what I do already exists in some form or another (fluid dynamics). But my job is more of understanding what I do so that I can convince someone else that my programs are robust and reliable. So again, I'm programming a lot with AI, but the important stuff, it's my skill to understand it, not AI.
What do you suggest these people do instead? I’ve been frustrated by this recently: I pivoted to a new subfield and am working on stuff that I would love to dive deep into and really learn what is going on so I can speak intelligently about the tradeoffs etc. But that would take a long time, and I have tasks that I should get done. So I have found myself working with a pretty vague understanding that, when pressed by coworkers, quickly finds its limits. Then I go back and try to deepen my understanding enough to cover those limits. But because I’m not working with each detail of the problem, implementing line by line with time to think about what’s happening, there just isn’t time for me to learn this unfamiliar topic. But I would love to, and I would enjoy the work much more if I could. So what am I supposed to do?
I'm not GP, but I've been experiencing this too. All my coworkers are in the same boat though, so we don't really have anyone with the deep understanding. Our approach (which has been working pretty well, we've solved a lot of problems and learned some along the way) has been to use AI as an informer who also points us to sources (like documentation, github issues, etc) where we can then read up on something the AI has pointed us at. We try to time box things a bit to prevent going down the rabbit hole, so we don't always get to exhaust our curiosity, but we're continually gaining that knowledge and pressing forward. Most recent example is tuning the kernel settings and our app on our prod machines to behave better for WebRTC packet handling/forwarding. AI for things like ffmpeg has been a god send when nobody on the team is an ffmpeg expert.
I've also had success just asking claude to write documentation on a system or subsystem or module, etc, and reading that. I then sometimes have it turn that into an svg diagram or something visual that often helps understand things. That domain-specific knowledge is hard to gain though, so not a silver bullet by any stretch.
I don't think you can do much about it if you're hoping to hide that you have been in the domain for less than a year. You're going to have to admit your inexperience and ask questions of your more experienced coworkers and try to stretch your common sense (e.g. devise internal consistency checks).
Don't just blindly pass around LLM generated content that you don't understand. I think it's important to consider that you are supposed to be more than just a meat proxy for the model. You don't want people to start associating you with AI slop.
Yes, certainly, in a year I’ll be better. But in the meantime…? 3 years ago it would be understandable for me to take my time with the tasks I’m doing because they are complex. Now it would be unreasonable for me to spend enough time to understand the nuances.
reply