Hacker News Reader

> I never worked in a large company

I don't want to be rude, but it sounds like you don't know what you're talking about.

It often goes far beyond "insurance." A lot of this stuff is performative. Some people simply aren't good at executing, and many managers are focused on appearances and increasing the size of their kingdoms. Most can't allow themselves to be fine setting a roadmap and having it execute smoothly. There are all sorts of other perverse incentives and absurd behaviors I could go into, but the notion that it's all completely rational is silly.

I believe (s/t)he(y) means in the sense of risk management. A lot of things are done to ensure that certain risks are minimized.

>> The funding bodies want us to detail all expenses and justify all the changes in the plan, because they don't want people to say that the public funds are mismanaged. This is just an example. We also write tons of useless papers, to justify that we are productive with this money.

See how the solution is causing a different form of the problem it's trying to solve?

I agree with your premise, but sometimes I think having multiple efforts in parallel and letting many fail might be more productive than going the safe route.

I could see that working for a set of moonshot efforts at startups who primarily only have to keep investors interested, or some subset of research projects at FAAMG companies with money to burn.

But how could that work with big companies making contracts with other big companies as their customers? Doesn’t seem like they could afford to just “let many fail” as they could lose the customer’s trust.

Big corps doing big projects are already doing risky stuff. Those big projects often fail to deliver the expected result within the expected time frame and budget.

Devils advocate: Having multiple efforts in parallel with the crazy overhead would probably give at least one of them a fighting chance of succeeding. Who cares about the ones that failed when the whole big project worked for once.

I see, we are referring to an internal mechanism for the vendor to try multiple routes to meet the customer ask - this makes more sense, but I could see a lot of traditional enterprise customers being hesitant to buy in if they were made aware (if you could do it with only the partial resources that delivered the “successful” solution, why am I paying for the others?)

Most software is essentially rented. It already written and has zero marginal cost. The vendor is using the rent to develop the next version. An exception would be specific contracted work, or bug fixes under a maintenance contract. How that next version gets developed is really up to the company.

Or we can just continue getting great stuff like windows 10 and 11.

Well the original article talks about how the work that a software company does in a year, can be done in a few months by a freelance developer with zero bureaucratic overhead. But which freelance developer? For every freelance developer that succeeds, there will probably be ten or a hundred that fail, or push out a mediocre product. This is what all the meetings and design docs and testing is for. To increase the chance of success.

I'm sure the same can be said for the academic world. If I was in academia and I was given funding without oversight, I would probably waste it all on researching things I found interesting but weren't necessarily useful to anybody. If you don't want to have to justify your expenses, then just fund it yourself.

> multiple efforts in parallel and letting many fail

I actually worked for a company that did this for a while and the results were amazing. There were multiple tasks, each task being a way to improve or cost-reduce a part of the machine we built. Each task had 3 teams attempting to solve it a different way. We met back 6 months later to compare/contrast the outcomes. I've never seen so much innovation in such a short time. The job was a blast for those 6 months because you were only limited by budget (we had to build hardware) and imagination.

And then, of course, since it was a multi-billion dollar company, the entire thing got shut down and we went back to business as usual. But at least I still remember it 10 years later :-)

I don't think it's about safety and stability at all, given how CEOs everywhere are _begging_ for their employees to innovate, be accountable and take risks. All this insurance is required because large organisations tend to devolve into "cover your ass" culture. People get automatically promoted by staying, so avoiding getting ousted by being someone's else scapegoat is the safest bet.

This. Incredibly risk-averse behaviors indicate an industry ripe for disruption.

>CEOs everywhere are _begging_ for their employees to innovate, be accountable and take risks

Executives like to say these things, but what they say and what they mean often diverge. The nail that sticks out gets beat down.

> provide insurance to the different stakeholders of the project

A good analogy. The lower trust the environment, the higher the levels of insurance required. But it's more than this, it's almost like the "busy work" is a form of "Proof of Trust", and the lower trust the environment the more Proof of Trust is required to be exhibited.

It does seem like that's how things are. I wonder if it's how things have always been? And if not, what's changed?

This is a very interesting questions. I don't have an answer.

Whilst I agree with a lot of what you say, this part:

> We also write tons of useless papers, to justify that we are productive with this money.

doesn't seem to be a logical part of your argument. Writing useless papers is not a form of insurance nor proof that work is being done. It's rather the end product, what all the 'insurance' is trying to ensure gets done. The fact that academia outputs a torrential flood of useless papers has little relevance to the corporate world where the final product is, hopefully, useful despite sometimes being inefficient to get there. After all companies that produce truly useless outputs do eventually get the rug pulled from underneath them, either by bankruptcy or forced sale.

Yes and no. A scientist should generate knowledge and insight, the scientist must discover things. The only way to prove that to a bureaucracy is through papers. It should not be the end product, but it has become the end product. In any case, it is still insurance, if you publish cited papers, you are safe, even if their contents are useless.