I redefined myself to be Usain Bolt, so now at least half of you are faster than Usain Bolt.

If the premise that the strong version is based on is found faulty when building a weaker version on the same premise, then that's a weakness.

There is a risk of a reductionist fallacy. For example, the Markov assumption is that you can analyze a cryptographic algorithm one round at a time independently. Because the previous round, if ideal, acts as a completely random input into the next round. In reality if Round 1 introduces some bias, that gets passed onto following rounds. There is also a risk of local versus global optima. You can analyze every S-box and mixing formula completely independently, and even build a formal proof that such a step is optimally secure in isolation. But that can fall apart because it does guarantee that the combined algorithms is simply a sum of its parts. A flaw in the interaction between stages can subtly create vulnerabilities that the original components do not posses.

Gender is a social construct, so studying it is equivalent to studying something imaginary? Is that the gist of your argument?

A hot take, but not entirely wrong.

Thankfully journalists, opinion commentators, and political pundits stick to fact-based reporting; and would never spin an issue with serious global consequences into another ideological identity as part of a false left-right political dichotomy.

Yes. Have you ever seen stores where everything is perpetually 30 to 50% off? Notice how furniture stores are always having a "liquidation sale" ?

What's wrong with passwords. I can type mine in while I'm driving. I sure as hell can't scan the phone 2 orbits around my head under ideal light conditions in order to satisfy the face/selfie unlock crap.

But honestly, why do we have accounts at all? 99% of the crap we do on our phone doesn't actually need it. Forcing people to log into web sites in order to post seems like a good idea, but really it's still full of bots and sock puppets. Tear off the bandaid and lets all just go fully anonymous. If I don't need to receive phone calls or emails on my smartphone, then let's not even bother with logins.

If companies are going to insist on tracking all my activity, I feel like we're pretty close to running AI agents sufficient to bullshit their way through registration pages, captchas, throw away SMS sites, throw away email sites, etc. How much effort remains to make an Android fork that is just a platform for anonymous browsers and apps, where I don't need to login, and if the border agent goes through it, it's just whatever trash is still in RAM since I last booted it. Hell, if someone steals it I'll have to buy a new phone, that's fine the screen on mine is cracked.

(I know, it's a fantasy. human beings want to hand all their personal information over to tech companies. It's like an addiction for most people)

A good 2/3rd of the Slack apps at my job are in-house ones. And I wrote a few for Discord for my own personal use, I have make 6 hours invested into that project. And that was without any AI. With something like Claude or Copilot, it's a 30 minute job.

Slack is going to die pretty quickly. I don't see how it will survive. I say this even though the Fortune 500 I work for uses it for everything and have almost entirely abandoned email. It's just too easy to replace now, even with our technical debt. Because again, I can just say "port this to that" and in a day the stupid plugins and bots I have that only act as glue to the actual back-end run on a new system that isn't Slack.

Not really. It's not like Company X running Service X for me is going to get a better rate for hosting than my own company can get. Why pay a third party to run software on another third party's computers? I can cut out at least one layer of middle-man and pay the CSP to host and I'll just point Ansible at it. And let AI update the playbooks.

Desktop apps. You mean where half my corporate apps are Electron and JavaScript ? Well maybe not Microsoft, but we just ditched Outlook this year. So even the MBAs are going to be using the same weird stuff that HW and SW engineering teams here.

I no longer believe that anyone has to do those things. Or rather, substantially fewer people are needed in order to do those things.

He has the same plan for America. He'll make a few billion and everyone else is going to pay.

That's why my Dad voted for him the first time. So run this country like Trump Entertainment Resorts or Trump University.

My family gifted me the board game Trump: The Game, back in the late 80's or early 90's. I played it five or six times. Not the worst board game. At least the components were cool, lots of plastic buildings and vaults where you could put secret bids into. Which I assume represented the illegal kickbacks and bribes necessary to operate a business according to Trump.

We're pretty close to a point where you can ask AI to make you the platform and tools you want and it just does that. Why would I pay for github or slack clones when it's been my experience that AI can craft a MVP in an hour long session and iterate on a more complete service as needed later.

Buying software is already dumb when we have open source. And now paying for a host of software is also going to be obsolete.

You should simultaneously believe both that nearly all AI investment is a bubble and also that AI tools are not going way.

At the risk of an appeal to authority fallacy. I do have some background, as I used to write drivers for two kinds of deep learning / DNN accelerators. One for inference that can slice and rotate memory views for TensorRT, and one for training that transparently distributes virtual memory on a multinode cluster.

We humans can easily understand an inference engine. It's not hard, and I would recommend checking one out, even if just from an introductory YouTube video. Or diving into something like llama.cpp is worth doing if you want to see the underlying math. The ugly bits are where it hooks into various backends, but there is a core piece that anyone with a decent C++ background can understand. Without C++ background, I'd recommend looking into simple-llm.

Starting with a simple but useful model like TinyLlama is a good point of research. Low requirements and the model is small enough to experiment on.

The models are a blob of data. Introspection is very limited, there is no history, and the weights (tensors really) are interdependent and even circular. So adjusting one weight can rippled through the model with unintended consequences. There is very little that one can formally reason about in a general way with a model. Maybe we'll have that kind of math one day, but right now we are stuck treating models as a black box.

So my response to anyone that claims that humans already understand what an LLM does is: Then show me.

That's perhaps the most interesting take on all of this. That's the psychology behind why we are running headlong into something so socially and economically disruptive that there are going to be more losers than winners.

And I get a lot of flack for talking about it. Probably because nobody likes being reminded that they have been fooled, conned into believing in the hype and dreams, while simultaneously ignoring the consequences and accountability. But at the end of the day regular folks are going to pay the price, because the investors and creators of technology will bear little practical responsibility for a problem this huge.

Nice thing about "drink 5 cups of coffee" is that anyone can do it, with really no effort. Even if the alleged benefits are just barely measurable with statistics.

Gen-X, but I can understand the confusion given that I can afford to own a home.