I never understood how anybody would think this way. At work for example from day one of using AI our rule was, AI is just a tool and if you break something due to not reviewing the code properly etc. it's on you as if you wrote that code yourself.
Nobody ACTUALLY thinks this way, they just want to break all the laws and make their billions without any consequences. We've already seen the big ai labs turn around and cry foul when others try to use their content but they are happy to continue doing it themselves.
The hypocrisy and willingness to play dumb are kind of staggering.
The company owners have one set of standards for those that work for them, and a completely different set of standards for themselves and for the company itself.
If you are using AI and it causes some damage to them, then it's your fault and you'll get reprimanded/fired. If they deploy AI and it causes damage to someone else, then that's proof that AI has extraordinary capabilities that no one could predict and that merit way more investment in this tech.
Well, responsibility laundering isn't exactly new. People use corporations to diffuse responsibility and avoid punishment for crimes all the time. "The corporation did it, not the decision makers in the corporation!"
But if you are in a full selfdriving car and it kills someone, is it on the user, or on the vendor of the car? Or does it depend on what is in the EULA and whether the user agreed to it?
This is a great rhetorical question. Can you believe that an LLM user is responsible for what the AI does but a self-driving car (SDC) user is not responsible for what the AI does? It doesn't seem very logically consistent.
In both cases, the company operating it promises they did lots of things to make it safe. SDC companies talk about testing, redundancy, simulations, and statistics. LLM companies talk about alignment, sandboxes, defense in depth, and restricting access to dangerous capabilities.
The one substantial difference I can see is that LLM chats are (nominally) passive but SDCs are active. In a chat, you're just having a conversation and then you're choosing how to act. When riding in a car, you the user are not in the loop between AI and taking action that affects the world around you. So they might be designed and engineered a bit differently. A SDC can't get you to agree to review the steering and braking decisions before they're put into action. Since you're not being asked to take that responsibility, it's more defensible to infer that the service has done whatever is necessary to make it safe.
The practical answer is you make sure damages are always covered by insurance, and as long as nobody acts maliciously or with extreme negligence there won't be any need for prison time.
Practical in what sense? This is exactly how you get EPA superfund sites while the executives that profited from the offending activity remain unaffected.
Practical in that we get to have widespread self driving cars. Real ones where you can look away from the road.
Do you have an "offending activity" in mind? The existence of car accidents isn't one. If you want the manufacturer to be liable for every mistake the car does, that's fine. But since that will increase the price of self-driving, likely as a subscription, it basically resolves into a type of insurance.
If the self-driving designers perform actual wrongdoing, that's a separate issue that can be handled like any other recallable hazard.
Its like companies handing over race cars to delivery drivers and telling them to drive responsibly but unofficially tell them that they expect reduced delivery times due to faster cars.
Do you think the average rate of accidents will stay the same ?
Weren't there multiple microcosms along similar lines in the US with "30 minutes or less" pizza and _[gestures vaguely in the direction of Jimmy Johns]_ ...
The efficiency of responsibility laundering depends on your social standing. If people like you they welcome reasons not to prosecute you, if you are disliked people might go as far as to start prosecuting you for made up reasons.
Institutional rules that require people to break them (eg putting pressure on them to work faster than possible while adhering to all rules) are a great means to selectively enforce them to get rid of undesirables.
We have ChatGPT Pro at work and I usually use Terra medium/high and only bring out Sol High when the big or feature actually requires "thinking"/complex behavior. This has worked pretty well for me and it's very token efficient
How much code are you shipping in a day? I find I can pretty comfortable use Sol high most of the day and stay within the 5 hour limit. I’ve got too many meetings to allow me time to write code continuously for an entire day. I usually finish about one ticket a day and then review 1-3 tickets for my colleagues.
We have a very lean development process and are in early stages of releasing the software, so almost no meetings and 90% of my time is development time (obv I talk with colleagues etc in that time). For example working with yocto eats through tokens like crazy since codex has to work with a pretty big codebase and look up tons of stuff.
I noticed the same. I wanted a simple crud webapp and suggested an insane techstack involving C#, Razor Pages, MSSQL and more. I went with my planned setup of python flask with an sqlite db which served me well for years.
It's still incredibly important to have a human in the loop correcting design decisions and having good taste.
Was your prompt just "I want a simple crud webapp" and that's the extent of it? There's absolutely no way you included the words "python", "flask", or "sqlite" and it still went with a Microsoft stack.
Imo a good example are soccer teams. To be in the top 5 the team has to work as one hivemind which only works if the people making up the team (coaches and players) actually work together on both the sport side as well as on the social side of things.
Different to corporate jobs those teams spend tremendous amounts on scouting, analysis etc.. They don't just look for the best player but also one who fits as a person.
I think the core issue is that management hasn't picked this up yet and understanding team chemistry etc is totally undervalued. I'd argue this is partly due to this being hard to measure and very hard to do.
I always suggest looking at the quantum eraser experiment to see how bad common sense works with quantum mechanics. It's fairly simple but intuition fails most people when it comes to explaining how it works.
No need for funky stuff, just do the classic (lol) Young's slits experiment which should be more than enough oddness for anyone.
I was shown it at school, using microscope glass slides that we waved over a Bunsen burner running cold, to coat with soot. We then carefully etched parallel lines with a compass by hand. Our (~17 y/o) efforts were a bit random but we did get some smudgy banding results on the screen.
It was the '80s so safety razor blades were available but in a posh public school (UK version) aged ~17, compasses were everywhere and large old style razor blades were not - the modern style razor with one or two thin blades was a thing. We could have probably got some needles from matron for a finer point, than a compass.
reply