Speed and bottlenecks

Software people can be annoying.

Maybe it’s (was) the absorbing nature of the work that causes them to be blind to what seems obvious in other practices.

Maybe outside of software, you get humbled faster and more viscerally by the more experienced.

Now that we have LLMs and on-the-fly generated bash scripts, our collective consciousness has adjusted to a new baseline. All of a sudden, not having ‘ai’ ‘agent’ output logs stream in a terminal 24/7, some feel like they are ‘falling behind’. Honestly, I’m not sure what that means.. you the psychosis process at any time? If this stuff is proto-AGI then surely we can just jump in at the last minute and get absolutely mind blasting psychosis, without suffering through the “this is just the beginning” phase? Fuck it, depending on your risk tolerance, and retirement account, you can nope out of ‘work’ and wait it out until the world of ‘abundance’. Elon, if you can spare a few hundred million to a stranger (since it won’t matter) I won’t deny the gesture!

Is taking an extra few minutes, hours, days, months even, to improve yourself, through deliberation, reading, hands on practice etc going to make you ‘fall behind’? This makes no sense. For endeavours of any significance, success is measured over a much longer time frame, and taking the time to do things deliberately compounds over time - the same is true in the opposite direction for doing things carelessly.

Other observations; Executing with speed + precision requires a lot of training, can be draining and can be dangerous.

It is usually better, when less experienced, and to keep your skills sharp, to practice slowly. Often it is harder (maybe on the ego) to practice slowly because it immediately exposes flaws. This is more the case for people with just enough experience to be dangerous.

This is bigger than software. The idea is that as ’technology’ ‘advances’, we can gain ’time’ back, and spend it on ‘better’ things. Usually, this technology will get cheaper over time, and the standard of living ‘raises’ for most.

Judging by the S&P 500 company list, those better things are: consuming advertising, ultraprocessed art and ultraprocessed food.

Also, over time the representations of societal status reverses.

It used to be that if you had time to philosophise, you were likely part of nobility, aristocracy etc. For the most part. Now, you’re a bum because you should be working.. doing business, philanthropy.. just not laying around doing ’nothing’. At least not in the US.

Being able to, and actually overeating used to indicate wealth, now it mostly represents the opposite. You can associate certain types of diets, and even BMI with social class.

Everything is a status signal.

The type of sport

The type of entertainment

How and where you travel to

Whether you hike or not

Soon building digital products is gonna be a poor man’s thing. Why would you waste time doing that, when you could be.. playing instruments all day? Let alone, building digital products in a now, artisanal way. Being close to the code and the product being built. You would have to be dead poor, or filthy rich to spend time doing something which could be done for you by someone/something else.

But in the meantime, I think the truth still remains. There seems to be a universal principle we think we’re getting away with violating. We may be on the verge of experiencing the biggest, longest, violation in modern history, maybe ever.

“Take shortcuts, you get cut short”

I believe that the best way to use LLMs is to improve your own reasoning capabilities. To enrich yourself. This is is at best, as effortful as not using LLMs, at worst, even more effortful because of the time compression effects. The better you are, the better you can wield the tool, but the genie tempts us with a lot more than 3 wishes. (Go read up about genies and 3 wishes)

I also believe that most people are lazy when it comes to things they don’t have a deep connection to. So, my bet is that, LLMs and agentic coding will be a net-negative to most people and definitely most organisations because you can’t violate the principle without consequences. Everything will happen faster, features, fixes, convergence of design and UX, bugs, burnout, apathy, boredom.. death