Hacker News Reader

Comment by valleyer | Hacker News Reader

> The first thing that blew my mind was how stupid the whole idea is. Think for one second. One full second. Why do you ever want to add a chatbot to a snack vending machine? The video states it clearly: the vending machine must be stocked by humans. Customers must order and take their snack by themselves. The AI has no value at all.

I fear the author has missed the point of the "Project Vend" experiments, the original write-ups of which are available here (and are, IMO, pretty level-headed about the whole thing):

https://www.anthropic.com/research/project-vend-1

https://www.anthropic.com/research/project-vend-2

The former contains a section titled "Why did you have an LLM run a small business?" that attempts to explain the motivation behind the experiment.

View on HN

Yeah, I haven't read the WSJ article, but I did read the original Anthropic experiment and I feel like the author is catastrophizing a bit much here. This is effectively just something they did for fun. It's entertaining and a funny read. Not everything has to be the end of the world.

The point is that it's an ad. No company spends money on a joke just to make a joke. Not the end of the world, although it's interesting that all the end of the world stuff comes directly out of that joke and its universe as it were. Take the joke seriously and extend its logic as far as it will go, and you get the end of the world. It's a thought experiment, or that's how I read it anyway.

I think the phrase we're looking for is "publicity stunt". Seems a fairly harmless and self-effacing one at that.

> The point is that it's an ad.

Sure, but like the other guy said, that's the point of publicity stunts. It doesn't even have to be specific to a company/ad, any silly thing like this is going to sound crazy if you take it seriously and "extend its logic as far as it will go". Like seeing the Sony bouncy balls rolling down the street ad and going "holy shit, these TV companies are going to ruin the world by dropping bouncy balls on all of us". It's a valid thought experiment, but kind of a strange thing to focus on so sternly when it's clearly not taking itself seriously, especially compared to all the real-world concerning uses of AI.

(And it is pretty funny, too. If anything I think we'd all prefer more creative ads like this and the bouncy ball one, and less AI-generated doomer Coke ads or such.)

To be fair they could have done an experiment on a transaction that typically requires a person in the loop. Rather than choosing a vending machine which, already, does not require a person in the loop for the transaction.

If I recall, the idea was the AI taking the role of the vending machine manager, choosing and restocking products and such. Anything on top of that was, I assume, just added for fun.

You’re thinking of replacing a vending machine with a chatbot which indeed doesn’t make much sense. The experiment was replacing the management of the machine. It’s not crazy to think that, money being no object, it would be great to have a person who hangs around the machine, whose job it is to ask customers what kinds of things people might want to buy from the machine, and conduct trials, tinker with pricing, do promotional deals, etc. But that’s of course impractical to have a FTE per machine to do that. The idea of this experiment was to see if that could be done with Claude. And of course as others have pointed out, it’s a simple and cheap and low-stakes version of “a business” suitable for experimenting with.

>I feel like the author is catastrophizing a bit much here.

I feel like he's catastrophizing the ordinary amount for an anti-AI screed. Probably well below what the market expects at this point. At this point you basically have to sound like Ed Zitron or David Gerard to stand out from the crowd.

AI is boiling the oceans, and you're worried about a vending machine?

The partner for these projects has a benchmark that the top frontier LLM labs seem to be running on their new model releases - I think there's _some_ value to these numbers in helping people compare and contrast model performance.

https://andonlabs.com/evals/vending-bench

Reply on news.ycombinator.com