Someone tried this, I saw it one of the Reddit AI subs. They were training a local model on whatever they could find that was written before $cutoffDate.
Agreed, I get better design and arch solutions from it. And part of my system prompt tells it to be an "aggressive critic" of everything, which is great -- sometimes its "critic's corner" piece of the response is more helpful/valuable than the 'normal' part of the response!
OP posited SQLite database contention. I don't know enough about this space to agree or disagree. It would be interesting, and perhaps illuminating, to perform a similar experiment with Postgres.
Yes, 100%. Being immobile is bad, whether you're standing, sitting, or lying down. (Too much strain from movement is bad too, hence RSI for many of us.)
I really enjoy having the flexibility/option of standing or sitting, and IMO a standing desk is one of those purchases that has a 10x payback.
Another comment mentioned using a linter to prevent 'console.log' from being mergeable in a PR, and this is exactly the kind of approach I'd take. Preventing an invalid state from existing is a pretty useful principle.
Even if one refuses to move on from software dev to something like AI deployer or AI validator or AI steerer, there might be a need.
If innovation ceases, then AI is king - push existing knowledge into your dataset, train, and exploit.
If innovation continues, there's always a gap. It takes time for a new thing to be made public "enough" for it to be ingested and synthesized. Who does this? Who finds the new knowledge?
Who creates the direction and asks the questions? Who determines what to build in the first place? Who synthesizes the daily experience of everyone around them to decide what tool needs to exist to make our lives easier? Maybe I'm grasping at straws here, but the world in which all scientific discovery, synthesis, direction and vision setting, etc, is determined by AI seems really far away when we talk about code generation and symbolic math manipulation.
These tools are self driving cars, and we're drivers of the software fleet. We need to embrace the fact that we might end up watching 10 cars self operate rather than driving one car, or maybe we're just setting destinations, but there simply isn't an absolutist zero sum game here unless all one thinks about is keeping the car on the road.
AND even if there were, repeating doom and feeling helpless is the last thing you want. Maybe it's not good truth that we can all adapt and should try, but it's certainly good policy.
> Maybe it's not good truth that we can all adapt and should try, but it's certainly good policy.
Are you a politician? That's fantastic neoliberal policy, "alternativlos" even, you can pretend that everybody can adapt the same way you told victims of your globalization policies "learn how to code". We still need at least a few people for this "direction and vision setting", so it would just be naive doomerism to feel pessimistic about AGI. General intelligence doesn't talk about jobs in general, what an absurd idea!
Making people feel hopeless is the last thing you want, especially when it's true, especially if you don't want them to fight for the dignity you will otherwise deny them once they become economically unviable human beings.
I think you jumped way past the information I shared. I don't think it's productive to lament, I think it's productive to find a way to change or take advantage of changes, vs fighting them - and that has nothing to do with globalization or economics or whatever, I'm thinking only about my own career.
I’m not sure I understand the point about learning. But wouldn’t any job that is largely text based at increased risk? I don’t think software development will be anywhere the last occupation to be severely impacted by AI
I don't know the first thing about such things, but I'll bet if companies were pushed on this they'd suddenly start asserting that it's a lease, which I think is typically taxed similarly to a sale.
I hate the "everything's a subscription" business model that's taking over everything. We'll achieve peak serfdom when the air we breathe, water we drink, and food we eat is bought on a subscription model.
Food sometimes is, and honestly a more centralised food subscription system could drive down the cost of food by making demand more predictable and enabling better economy of scale.
If it were feasible to charge you for your own heartbeats, companies absolutely would. Infinite growth is a poison pill idea - it brought us to where we are today, and would drag us to the future you outlined, unless the folly is abandoned.
Nothing wrong with growth,it's natural. The population grows and so does the money supply in turn. This is good. What is not good is when a particular company is trying to retain ALL growth of certain segments for themselves through abusive leverage.
A principle that stuck out to me as a child was that our society prioritized lower prices for consumers as a whole over the prosperity of any one company or industry.
We let companies grow to the point where they now subvert the will of the people.
Growth is good until it's not, cancer would like a word.
The only thing that seems capable of growing forever is the amount of space in the universe. Line go up increasingly fast always forever is a bullshit fever dream that needs to be culled with prejudice.
Not in real estate. There is technically no difference between leasing and renting, other than the things that are colloquially associated with. A lease agreement (or rental agreement) may or may not have a clause offering the borrower an opportunity to take ownership.
Rent might also refer to the price portion of a leasing agreement (or rental agreement). So in shorthand usage, “the lease” might encompass all terms of borrowing something, including the price, but “the rent” might just refer to the price.
My fear/anxiety causes me to avoid action. Fear of mistakes, failure, reproach/criticism from others, that it's going to take a lot of time/money, that I'll waste a lot of time/money, etc.
Then, once I've procrastinated, my fear/anxiousness of the consequences of NOT doing the thing kicks in and I rush to complete it.
It's great! /s
The best solution I've found is to slow down, introspect into and address my fear/anxiety, and reassure myself that most of the time my fears are entirely or mostly unfounded, and that even if something "bad" happens it won't be catastrophic.
I don't practice that enough, and that's the hill I'm climbing now. I'm also trying to learn to lean more into the feeling of accomplishment ahead of time. In a "it will feel so good to get this done, even if it's not great" way.
I think the only way this gets better is with software development tools that make it impossible to create invalid states.
In the physical world, when we build something complex like a car engine, a microprocessor, or bookcase, the laws of physics guide us and help prevent invalid states. Not all of them -- an upside down bookcase still works -- but a lot of them.
Of course, part of the problem is that when we build the software equivalent of an upside down bookcase, we 'patch' it by creating trim and shims to make it look better and more structurally sound instead of tossing it and making another one the right way.
But mostly, we write software in a way that allows for a ton of incorrect states. As a trivial example, expressing a person's age as an 'int', allowing for negative numbers. As a more complicated example, allowing for setting a coupon's redemption date when it has not yet been clipped.
John Backus's Turing Award lecture meditated on this idea, and concluded that the best way to do this at scale is to simply minimize the creation of states in the first place, and be careful and thoughtful about where and how we create the states that can't be avoided.
I would argue that that's actually a better guide to how we manage complexity in the physical world. Mechanical engineers generally like to minimize the number of moving parts in a system. When they can't avoid moving parts, they tend to fixate on them, and put a lot of effort into creating linkages and failsafes to try to prevent them from interacting in catastrophic ways.
The software engineering way would be to create extra moving parts just because complicated things make us feel smart, and deal with potential adverse interactions among them by posting signs that say "Careful, now!" without clearly explaining what the reader is supposed to be careful of. 50 years later, people who try to stick to the (very sound!) principles that Backus proposed are still regularly dismissed as being hipsters and pedants.
I'd say that the extra moving parts are there in most cases not because someone wanted to "feel smart" (not that it doesn't happen), but to make the pre-existing moving parts do something that they weren't originally supposed to do, because nobody understands how those pre-existing parts work well enough to re-engineer them properly on the schedule that they are given. We are an industry that builds bridges out of matchsticks, duck tape, and glue, and many of our processes are basically about how to make the result of that "good enough".
Found the GitHub: https://github.com/haykgrigo3/TimeCapsuleLLM