Mika

https://youtu.be/4Xkw8ip43Vk

Can you elaborate on infrastructure part? I thought dev efforts are the key issue?

That’s a good value! I was worrying something running mid power LLMs would cost 10k+ today. Sure ram screwup nuked the field, but it’s also not on the top list of my priorities. So it gives me hope that by the time I would be ready it would be in even better shape.

If you are capable of running 120b, the mistral cli bigger open source model in agentic mode could be available to you.

How much did your setup cost, if you don’t mind to disclose such information?

Not really. But the more you understand tech, the less you like the high tech market solutions and have the urge to have the same but in opensource.

I kinda dream that one day setups for mid-power home AI racks would be affordable enough to drop all the proprietary AIs and start automating my daily routines with agents & helping me build my local knowledge vector database with AI embeddings. No way in hell I’m exposing anything like that to thirdparties.

I once was helping to organize the testing of town-level algorithmic competition for school students.

The competition had one entry level issue that was basically solvable by reading the question properly, recognising that it’s just multiplication of two numbers, and writing the simplest app ever.

And there was one student who passed the automatic tests. We had to read the code too for the protocol, just to make sure there was no cheating.

We looked in the code. What? Why? It had two nested for loops and a++ inside. When we understood what’s going on we couldn’t stop laughing for like solid ten minutes.

I mean that’s the whole idea, you don’t create knowledge out of nothing. But it’s more than just preexisting belief, you try to grab it from collective unconscious.

Plus, card on itself doesn’t tell you anything SPECIFIC. The whole point of cards and magic is to believe that this specific card has bigger meaning, thus you need to interpret it, and think about it’s meaning. And thinking is useful. What conclusion you come up with depends more on the situation than the card itself.

Unless you happen to live next to russia.

It’s in hundreds of millions and dude has like half a trillion. I’d say multiply by arbitrary number between 100 and 1000.

They already installed loyal government in USA and are trying to do the same with the key EU players. Ukraine will lose without support. Support costs a lot. It grows unrest that ends up in russian-funded far-rights getting to power.

Democracy is a double-edged sword.

I’m kinda considering. Cause if I’d build a custom PC, it could be a home server of sorts - some webserver or AI machine.

I have only certain amount of time I can spend on personal tech activity and if I could play without spending that energy that would be fantastic. Also that’s decent PC for Godot gamedev right?

So do it? Honestly the biggest issue is privacy, someone could analyse your dark fantasies in detail.

And yeah I think we will be able to simulate human brain at some point. What’s more interesting, would we be able to continue the thought process outside the biological box. Could I escape into machine? Doesn’t guarantee huge longevity boost as machines bug out & break from time to time, but still.

Consciousness is a process, not just data. If they copy the weights of your brain neurons, they would be able to create new life that thinks like you, but is NOT you. Although that new life would be very confused and would honestly believe that it is you.

Thus, it’s neither eternal life nor eternal torment. Maybe torment of newly created AIs, but we didn’t come to “AI have rights” yet.

you’re a code monkey

You write your code manually and I set up infrastructure to avoid that, who of us both is a code monkey? :-)

I write tech plans too, from time to time.

concrete example

I already gave one. With jira & figma MCPs you just tell “read ticket <link>, read figma, make a separate component named XYZ. Look at file QWE to follow the same code style.”

Yeah you can verify yourself, it’s still there.

Actually if you think about it, VK link in accounts on Instagram is a dead giveaway even without twitter, this is just an extra verification.

The best one in this series:

I regularly filter out tabs on laptop, bookmarking things and closing stuff that isn’t on todo list for the next 24h.

Mobile though, clicling that through UI takes so much time I can’t be bothered. I just open new stuff on top, and maybe sometimes go through tabs like as if that’s browser history.

This is literally in this thread.

Again, your solution should be already thought out and described in tickets and approved tech plan. If it’s not, SDLC problem.

And it’s not true that agents can’t help with edge cases, they can. If you know which points to look at, you task to analyze the specific interaction and watch which parts of the code would be mentioned.

I do write way less amount of symbols to LLM than I would when I write code. Those symbols don’t have to be structured and they can even have typos, so I can focus my brain activity on things that actually matter.

Plus, copilot is shit.

I rate your post as a skill issue.

This is necessary when working with 100% human code too.

deleted by creator