little rant

Its a valid rant.

There are issues with AI involving licensing and peoples IP and crediting sources and all that.

It has also massively reduced the barrier of entry for both finding and patching bugs, and producing realistic looking bug reports.
My biggest issue (other than enviromental and social impacts) is the load it puts on developers. They pretty much HAVE to use LLMs to triage and test bug reports… because it takes very little skill to use an LLM to “deep dive this area of the linux kernel and find a bug or RCE/CVE”, and the LLM probably will. But the unskilled user might not understand what the exploit/bug actually is. But they get some sort of POC and some sort of write up, and then they dump it on someone elses plate.
If that bug is actually patched/impossible because of something outside of the scope/context that the LLM used, or if its only a bug because of bad config… yeh, its a bug, but it isnt an important one. And it eats up developer resources to confirm/triage/fix that bug.

Also, LLMs are an averaging engine. They work on “most likely” probability evaluations. So, you get average code.

There has to be a human in the loop

There really is an XKCD for everything. Even if it was drawn years before what its applicable to

Purr I do not alter it furrier.

Hmm, maybe thats too much

I consider aetheism to be more an absence of belief in gods.
Its not that i believe there arent gods. Its that i have no belief in gods.
The former being anti-god/anti-religion. And the later being anti-belief.
Maybe like “not accidentally cutting yourself” and “not having a knife”. Both lead to “not accidentally cutting yourself”, but 1 involves a knife.

Yeh, this is common.
I’ve heard of PPI (Producer Pleaser Indicator) which is a mixing desk channel not routed anywhere but can be fiddled with to meet client demands.
And there is the DFA (Does Fuck All). Similar concept

You’ve never experienced an AWS US East outage, huh?

First of all, I like the alliteration of examples I always struggle with things like that, and alliteration is an obvious (now, to me) bridge for these.

explore ideas too large to fit in our brains

Another excellent explanation of programming.
And it really does describe the programming of tasks, from punch card weaving machines, to punch card data entry, to direct data entry, to punch card processing, direct processing, and eventually the whole Operating System, Assembly, and Higher Level Languages… And all of the steps between that have since developed.
You can become an expert in memory management, and that idea fits into your brain. And you build the APIs that mean memory management doesn’t have to fit into other people’s brains. So they can work on instruction optimisation, or storage optimisation.

Standing on the shoulders of giants.
There is so much that can be solved with programming these days, but actual people have so much more… Peoplesness, humanity, realness… That some things aren’t worth programming yet… That problem hasn’t been solved yet

It’s almost anything.

Objects? Python macro into FreeCAD into 3d printer. (Maybe there is a better way? I like having that final 3d editing space to tweak).

Literally any data-> data system. If you have structured data, and something wants structured data? Programmable.

Music is sooo mathematical when you get into it. But the creativity is recognising moving from one math to another, or just throwing math out the window. So yeh, programmable. Making it good? Creative

I hate to say, but LLMs show that some level of human interaction can be maths (it’s all just floating point multiplication). (And I say “interaction”, it’s not intelligent or creative or anything actually human).

If it’s maths, it can be programmed.

… It’s not “putting pixels next to eachother”. It’s figuring out how to filter noise to get it closer to what the prompt describes. Literally starts as noise, like an old TV that isn’t tuned to a channel. And then it applies a filter that it calculates will best get that noise closer to the tokens of the prompt. And then does it again and again and again

Regardless! Even if that investment went into healthcare, or a universal basic income…
Maybe we invest it into robotics to automate stuff. Or maybe we invest it into people.
Investing in people would be neat

But that will never happen until it is required to prevent collapse

Or you tell it not to commit but to suggest the commit message

What if… the government set expected prices for things? Say they set the price for 90% of treatments. That includes life long medications and multiple tentative surgeries.
If it’s on the government register, the cost is known.

Like, they say that 90% of medical treatment will be a fixed cost. And it starts off fairly expensive. But with aggregate data it gets cheaper. And the fact that the government mandates it. And I’m sure there is already a lot of data…

It costs 1$ for a dose of paracetamol (or whatever). Starts off whatever the average is, then trends towards the actual cost to acquire. Takes a year or 2 to bring costs inline, paracetamol is $20 per dose but 20¢ per dose, etc.

I mean, who the fuck is licencing these hospitals to charge this much? The government, right?
Who the fuck is licencing these pharma companies to sell the drugs this expensive? The government, right?
And who is letting medical insurance companies this much power? The government!

Why the fuck doesn’t the government actually align price with cost?
Oh, money. Yeh, it’s money.
It’s about senators being bought for $10k.
It’s about “lobbying” $10m, and generating a $1bn gap.

git is my backup

Git is just a folder

Pretty much where I am at.
Piracy is a service issue, I think Steam proves that.
If that service isn’t viable any longer, time to move on

I seem to remember a lot of arguments when Netflix first introduced a cheaper-with-ads subscription that “not everyone can afford ad-free subscriptions” and that somehow it would make things better.

Ignoring the fact that the play was obviously to get the ad-subscription up to the current cost of ad-free subscription.

And now I think it’s Disney introducing ads to all subscriptions.
So, no doubt Netflix will do that soon in order to “innovate” the line-go-up.

Humans ARE the resource

Just 2 alphas doing alpha things

Of course Brits don’t want to become a minority in their home country… who would? We all know how badly they are treated

Absolutely wild comment.

I’d rather have a more ambiguous name that is clarified by context than comments or docs.
create(obj) is useless, but if it is namespaces and properly typed user.create(User user) makes a LOT more sense.
Granted, that is a very narrow example, but it gets the point across: let context carry context, and use it properly.
Something like import {create} from 'user' (and using it with merely create()) is less useful that import * from 'user' (and using it as user.create())

We namespace things to help split up code. Lean on the namespaces when using that code!

Usama Arif found that with the three patches to just probe BMI2 CPU support only once, it yield a 71% reduction in decompression time and a 18% reduction in compression time for a crypto_acomp benchmark.

Wow