AI Is Making Us Work More

I'm not doubting you, btw... I've seen others here on HN also saying that they burn through money with AI, I guess I'm just missing something.

In fact, the geek in me absolutely wants to know what's going on, because you have probably found something that I would love to know about! :)

AI is currently making energy more expensive. Shelter and these other commodities aren't made cheaper if population expands along with any capacity increases (like lanes on a highway). Lots of "ifs" in this statement that don't seem to match with observation of reality. The point of the discussion here is that AI in many ways is making workers less efficient.

"If", of course, but it's all but certain. The energy situation, today, is that we as a species depend on fossil fuels, and they are depleting. And we don't have a solution to replace them (renewables are not remotely replacing oil today, those who believe it say something like "we've done 1% of the job, it proves that we will reach 100%!").

AI is making us use more energy, and our use of energy is what is killing the world. Again, the consequences of abundant energy are global warming, mass extinction and political instability as the fossil fuels get less and less available.

One could simply work less. Even if you're full in the office, you can just use that extra time to learn something new.

So, fraud? If you put a fixed price based on 3 hours, that's of course fine. If you lie about how long work takes you, that's fraud.

Unless your bids are what you bill not what it takes, and you would bill the same 3 hours if it took you 4. In which case it's a fixed price under a different name.

Initial build

database integration

accessibility

speed

But such is that state of the UK that I had simply assumed the government had censored it. Remarkable how quickly expectations have shifted.

Is this really how professionals work on such a problem today?

The times I'd had a tune the responses, we'd gather bad/good examples, chuck it into a .csv/directory, then create an automated pipeline to give us a percentage of success rate for what we expect, then start tuning the prompt, parameters for inference and other things in an automated manner. As we discover more bad cases, add them to the testing pipeline.

Only if it was something that was very wrong would you reach for model re-training or fine-tuning, or when you know up front the model wouldn't be up for the exact task you have in mind.

We've kept the LLM constrained to just extracting values with context, and we show the values to end-users in a review UI that shows the source doc and allows them to navigate to exactly the place the doc where a given value was extracted. These are mostly numbers but occasionally the LLM needs to do a bit of reasoning to determine a value (e.g., is this X, Y or Z type of transaction where the exact words X, Y or X will not necessarily appear). Any calculations that can be performed deterministically are done in a later step using a very detailed, domain specific financial model.

This is not a chatbot or other crap shoehorned into the app. Users are very excited about this - it automates painful data entry and allows them to check the source - which they actually do, because they understand the cost of getting the numbers wrong.

I think individuals who get comfortable in their jobs don’t like automation arriving at their station because it upends the order of things just as they were feeling comfortable and stable. Being adaptable now is more important than ever.

> Products don't get better either, but that's more of a "shareholder value" problem than it is a specific technology problem.

This is broadly false. Your laptop is unquestionably better because it was constructed with the help of automated CNC machines and PCB assembly as opposed to workers manually populating PCBs.

Some companies can try to use automation to stay in place with lower headcount, but they’ll be left behind by competition that uses automation to move forward. Once that leap happens it becomes accepted as the new normal, so it never feels like automation is making changes.

Now, if what you actually want is to be relatively more prosperous and have more status that's a game you can keep playing forever. But you really don't have to, to simply be better off than all people in the past with far less work.

For instance, I had to rename a collection of files almost following a pattern. I know that there are apps that do this and normally I’d reach for the Perl-based rename script. But I do it so irregularly that I have to install it every time, figure out how I can do a dry run first, etc. Meanwhile, with the Raycast AI integration that also supports Finder, I did it in the 10-15 seconds that it took to type the prompt.

There are a lot of tasks that you do not do often enough to commit them fully to memory, but every time you do them it takes a lot of time. LLM-based automation really speeds up these tasks. Similar for refactors that an IDE or language server cannot do, some kinds of scripts etc.

On the other hand LLMs constantly mess up some algorithms and data structures, so I simply do not let LLMs touch certain code.

It’s all about getting a feeling for the right opportunities. As with any tool.

Too many people are trying to jump to the end when they don't even have their day to day managed or efficient today can tend to carry forward efficiency in a number of business workflows.

Checking the LLM output is required when it's not consistent, in many cases maintaining the benefit requires the human to know more on the subject than the LLM.

Without automation we would all be living in poverty.

There are definitely many things which when automated loses out on some edge cases. But most folks don't need artisanal soap.