I suspect it depends on the company — eg, at Amazon the tech team went out of our way to blind ourselves to what the tax analysts were doing; at a fintech startup, we knew intimate details of what clients were doing.
This sounds like the kind of big-co where intentional segmentation and blinding would happen.
I used to write tests for a cryptocurrency trading platform. I regularly talked to the auditors, as they were the ones who made sure that the crypto we transferred into external wallets made it back into internal wallets after the test.
I wouldn't say that I knew a bunch of accounting practices from talking to them, but I _did_ learn that the CTO and director of QA both lied to me during my interview. Sure, we made _some_ money from the spread as claimed in my interview, but in truth, the bulk of our money came from loaning our customer deposits to 3 Arrows Capital. I knew we were fucked _months_ before the company suddenly went under.
Yup. Engineers do have visibility into fintech systems that they implement, maybe even more so than the bean counters since they can trace exactly which txns went where. These things are logged.
> Anything that removes the power of CEOs and gives it to the worker should be highly encouraged.
The only thing that will do this is if workers are the resource bottleneck.
> Economic democracy is the final frontier of human empowerment and giving workers then means to have democratic control over the economy can only unlock more human potential, not less.
This already exists. It's called free enterprise and freedom of association.
Unless of course you mean that nobody can own or expend resources without (nominally) everybody agreeing... which has also been tried, and failed horribly.
For one such example, see the years long fights in city halls over resource usage or utilization, such as building new developments for example. A corporation trying to get something done moving at that pace would, well, not get anything done. That is why worker owned co-ops, which you can create today even in this capitalist system we have, do not outcompete capitalist structures generally speaking.
> Do those CEOs think their job is more complex than a software developer job, which they are so eager to replace?
You can estimate the difficulty of a job by what fraction is the population can successfully do it and how much special training this takes. Both of which are reflected in the supply curve for labor for that job.
> How many times more urgently do we want to replace the CEO, considering salaries? How about we put as many times the amount of money into that, as we are putting into trying to replace developers?
Pretty sure that (avg developer pay * number of developers) is a lot more that (avg ceo pay * number of ceos).
Since businesses need to start somewhere/when and most startups fail, I think most people who even get into the role of CEO, are doing it successfully. However, this is a lot due to circumstances and many factors outside of their control. There are also many CEOs ruining their businesses with bad decisions. It is not certain, that an "AI" wouldn't do at least as good as those failing CEOs. Similarly, many developers ruin things they touch, introducing tons of complexity, dependencies and breaking user workflows or making workflows cumbersome without listening to user feedback and so on.
In short many people do a bad job and businesses are carried by others, who do a good enough job to make a net positive for the final product. Or consequences of messing up are happening slowly, like a slow user drain, or a user replacement with bad actors until good actors start to leave, or any other possibility.
About the pay argument: Well, these days you still need a good crew of developers to make the shiny AI toys do what you want them to do, so you are not replacing all of the developers, so you can't calculate like that. If we calculate some Silicon Valley CEO making 2 million and a developer making 100k-200k, then we are still at a ratio of 10x-20x. If we manage to make only one CEO obsolete or 2 out of 3 CEOs 1.5x as efficient, we have achieved a cost saving of 10-20 developers! Yay!...
Maybe first get the automated vending machines to work properly, and then revisit this question to find what tasks that allows to be peeled away from the role?
Vending machines are static experiences. They sit there and wait until told to give an item and paid for it. Why would you need an LLM for that? There’s nothing to solve there
The comment is a reference to a well known benchmark in which LLMs inevitably lose money attempting to run a simple simulated vending machine business.
You know those puff pieces are to assuage our fears of job loss and provide an folksy aww-sucks patina to the plucky LLM who can't run a vending machine.
I'm not sure you can get around the principal-agent problem that easily. Who sets the policy levers on the automation and governs it? They inherit the ceo's negotiating leverage with shareholders.
It seems like you'd need some sort of fairly radical control structure (say, no board, just ai interacting directly with shareholders) to get around this. But even this ignores that the automation is not neutral, it is provided by actors with incentives.
They measure/model aboveground biomass, and present the change in that measurement as being a source/sink in the carbon cycle, ie as coming from / going to the atmosphere.
But I also see multiple places they mention the changes as being at least partly due to logging or wood harvesting. Which seems like biomass being removed and yet not going into the atmosphere.
Prior to the recent RAM insanity(a big caveat I know) a 1u supermicro machine with 768GB some NVME storage and twin 32 core Epyc 9004s was ~12K USD. You can get 3 of those and and some redundant 10G network infra(people are literally throwing this out) for < 40k. Then you just have to find a rack/internet connection to put them in which would be a few hundred a month.
The reality is most sites don't need multi region setups, they have very predicable load and 3 of those machines would be massive overkill for many. A lot of people like to think they will lose millions per second of down time, and some sites certainly do but most wont.
All of this of course would be using new stuff. If you wanted to use used stuff the most cost effective are the 5 year old second gen xeon scalables that are being dumped by cloud providers. Those are more than enough compute for most they are just really thirsty so you will pay with the power bill.
This of course is predicated on assumption you have the skill set to support these machines and that is increasingly becoming less common though as successful companies that started in the last 10 years are starting to do more "hybrid cloud" it is starting to come back around.
They just happen to have an online config tool that is somewhat close to what you would pay if you didnt engage with sales which is useful for a hackernews comment.
I'm also not used to developers having that much visibility into accounting practices. And that seems like an odd way to structure things.
reply