The End of Rationing(sanketsaurav.com) |
The End of Rationing(sanketsaurav.com) |
The answer has nothing to do with rationing. When all software is designed to grab the most attention, serve the most adds, collect the most data, deliver the maximum amount of the cheapest content and generally be as exploitive of its users as possible, no amount of rationing or not rationing will effect it.
McMaster Carr, a company worth a little over a billion, sells metal, bolts, casters, and many other industrial bits and bobs. Their website is infinitely better than Amazons. Why? It has nothing to do with rationing and everything to do with trying to actual deliver a good customer experience.
Two factors:
- middle management do not have contact with either customers or shareholders. They report to managers, and only have managers reporting to them. They're insulated from the actual consequences of their decisions.
- the only way to get recognised for promotion is to create change. Any change. Doesn't matter if it's bad (blame can be shifted), it just needs to be noticed.
If a feature of the product hasn't changed for a long time (because it works for the user), it becomes more of a target for changing. If it changed recently, it creates less impact if it changes again, so isn't a good target. So the product becomes worse over time, not better.
Though, maybe the irony only exists because we use the broad term tech company instead of the more generally accurate terms data harvesting company and attention consuming company.
Their target market is people who need to fix something quickly or get some industrial supply item tomorrow so work isn't stopped.
Massive corporations save money via the end user by taking a chainsaw to customer service. Large tech companies (and many other industries) will just quietly fuck the regular customer over because they can't profitably do customer service at scale. Even Amazon's effectively no questions asked return policy is a band-aid over the forgeries and trash that fill their hyperscale market.
If you have a problem with an item that Mcmaster sends you, they're going to go have a stern talk with the manufacturer. They sell reliability, not race to the bottom capitalism.
(I've also heard that they really take care of their employees. People in jobs that are automated away are always retrained; they'll pay for college with no qualifications, because education is good even if they don't directly benefit, etc.)
It was much the same with search. Google search allowed us to find things faster at first...but over time websites responded to that and started doing SEO crap. Now almost every search results in a billion websites that say the same thing, but shallowly. Yes you can still find some gems but was it worth it?
AI makes things faster, but now it means we can target superificial needs very quickly...and so a new breed of crap software will be built with it. As long as there's a dollar sign at the end of the road.
I recall something I once heard about product marketing. The experiment was a product like jam, and they had 3 flavour options available in one case and 12 options available in another. It was a test on how hard it was to make a decision as well as how confident the person was they made the right decision later on.
One of Sartre's points on existentialism was that radical freedom leads to anxiety.
In some sense, AI is removing scarcity. It isn't clear to me that this is the same kind of thing as going from starvation to surplus. It is more like the jam case where we aren't increasing abundance in a total sense, we increasing the options within the existing surplus.
The trailblazer was Microsoft Office.
I was discussing this with someone else and used the analogy of nuclear energy. Although nuclear plants are highly complex and use forces that have only been understood for the last hundred years, at the end of the day it's still boiling water to create steam to spin a turbine, and nothing about nuclear innovation can engineer away the laws of thermodynamics.
Agents have and will continue to drastically change how we develop code, but it's my belief that drastically changing our software practices to speed up development just means we're going to hit engineering limits faster, rather than being able to overcome those limits entirely.
Unless, or perhaps until, we figure out how to use LLMs effectively at the organization level we're still going to be limited to the "slime mold" level of organizational intelligence.
Will that happen? Most likely, but will it take a generation or ten? I suspect it'll be far harder for society to adapt at that level. We've had Mythical Man Month for decades yet software managers continue repeating the same mistakes.
Because of that the idea of OAI and Anthropic just moving to in-house product development makes sense.
I’m not sure if the job title tho. Software restoration?
During the hype cycle, making "x for y" is easiest to get funded and exit.
Once the hype crashes and we're in the trough of disillusionment, then people are bearish. So founders either move to the next hype cycle or figure out a more durable business from the ashes
But if the answer to "hey Claude, I need to track my gym sessions, what's the best way of doing that?" is "I've built an app for your phone to do that", does that change?
To me, so far it seems like it is and it isn’t simultaneously. Some things are going at insane paces. I see things like reverse engineering and emulation breaking barriers. But we don’t see things like the quality and complexity of highly architected systems getting much better, maybe even getting worse as management pushes velocity onto employees.
I think what this shows is that AI is great at tedium and brute forcing, but it’s not actually good at making substantially better software with less of that “rationed” expertise. At least, not yet.
Tell me I'm wrong.
In the past, if you wanted to do a task, you had to choose from one of a few (or maybe just one) incumbents and make their tool fit your task. Failing that, build your own tool, but that was expensive. If you did build your own, the tool might have value and could potentially be sold to others.
Going forward, building your own will be much easier and cheaper. Even if the incumbent tools suit the task well, you might choose to build your own for a variety of other reasons. e.g. If the incumbent raises prices or chooses to enshitify. What you will probably not do is adopt some third party's vibe-coded clone of an incumbent's tool, because that won't be customized to your needs and it won't have the track record of the incumbent. You can't trust it or control it so it will be better to build your own.
If an incumbent exists and you can now vibe-code a clone, you probably can't sell that clone "as is" and expect to be successful. You might be able to sell your services to customize and refine tools for small businesses that can't really afford to hire a solitary soul to build their own solution, but don't expect that niche to persist long-term.
There are only so many rust migrations you can do, at some point you gotta ship something user visible which means you have to deal with people which means the code was never the bottleneck even in 2014
Now that SaaS CRUD apps can be built in 10 minutes, we can maybe do something more worthwhile, no? So much wasted effort has gone into this field already, in my opinion.
It's a different view of scaling. I mean a lot of people have this absurd idea of "super intelligence" which is like the cover of this book
https://en.wikipedia.org/wiki/Supermind_(novel)
i mean what does "intelligence quotient 10,000" even mean? On the other hand it's very easy to imagine a world where we have systems like the best ones we have now but we can really afford to use them.
IQ tests measure, essentially, how fast you can do shallow pattern recognition.
And AI can do these kinds of tests easily.
In Claude we trust.
There's no reason to think that companies are suddenly going to be spending 1:1 features : quality because both are faster / cheaper .
So expect the low quality to continue, and naturally get worse, because complexity is going up 1000x
And of course the number of libraries, known algorithms, etc, is much larger than 8. There's any number of ways to make something new by combining well-understood parts.
I don't think non-expert prompting is pushing the envelope on what can be built because I think the current generation of harnesses and models pushes to using well-understood patterns in well-understood ways but it's certainly moving the line for what has been built. And with more skilled prompting you can build pretty crazy stuff + extensive validation for it.
https://www.science.org/content/article/openai-just-changed-...