Alexey Skobkin

@strizhechenko Да похуй, вливай!

@pelle

sure, in a hypothetical world where we would have UBI, and data-centers were nice instead

Yes, we live in a real world, but that doesn't mean that we should not develop technology because it could be used for bad stuff or in a bad way too. This won't achieve anything just because this technology will be developed anyway. If not by your country then by another.
What could be productive is applying pressure to companies who do the harm without blaming the technology itself.

to such an extend that major NGOs call for a ban

NGOs also call for ban of nuclear power in some countries despite it being the most ecological way with high output available to humanity at the moment. The only fact that someone calls for something doesn't mean that it's justified. If I'd be using marginal examples I could've remembered people like Eliezer Yudkowsky who suggested bombing datacenters to control AI development. But that'd be a bad argument.

yes, in a hypothetical world where the generative AI provide benefitial medical advice instead of manipulating people to killing themselves
in our hypothetical world where LLMs would produce useful science instead of hallucinated nonsense

Okay. Here comes manipulation. I'm not sure if you aware of that or was it intended.

If it goes this way when we just ignore everything positive and exaggerate negative then I don't think there's a point in continuing this.
But thanks for a civilized discussion anyway.

@frssoft Нет.

@pelle

proponents of rifles would say

I don't care what would they say since this is irrelevant.

I'm more interested in what you would honestly say if I ask you whats the main purpose of an "AI" is. Your answer can help me decide if there's a point in continuing trying to argue or not. Some of your analogies and arguments look strange to me, but you're still communicating with me peacefully and didn't wish me to die, so I'm not sure which case is it.

Just in case, when I say "main purpose of an AI" I mean the purpose for scientists and researchers who're actually working on it, not VC's. We both know what VC's and corporations want.

mass-killing of pigs

You sure love to choose interesting analogies. Care to explain how is it better and outline the points where it matches the topic of our discussion?

your approach

My approach is a bit more complex. If you're really interested, I can say a bit more:

• yes, I think it's inevitable and wasting resources on stuff like blaming "AI" for everything is counterproductive. I'd prefer to spend them on researching of how we could apply it correctly, efficiently and how to help our society to adapt to these changes with less struggle. It may require introducing new laws like adding UBI or purging old ones like some copyright laws, et cetera. Also making AI open and available to general public, not only to those who can build datacenters.
• no resignation here. At least if I didn't misunderstand you - then sorry, English is not my native language. I'm inspired and optimistic. I look at what "AI" can do in science, medicine, how it enables people to do stuff they wasn't able to do before and look forward to new applications. Although it would be unfair to say that I don't see downsides of such a powerful tool being used for stuff like fraud and other crimes. I don't like how companies in their pursuit of maximizing profits can hurt people in the process - like building datacenters somewhere where people live. How they already caused price changes for hardware and basically made some people unable to buy a decent PC. I just don't mix everything up and don't concentrate on one single point like "AI is evil" and "AI industry is pure good". I look at it from the both sides and think that in the end humanity will benefit from it if it'll be smart enough.
• yes, I'll scold anyone for hysteria. I'll scold "AI" evangelists for blessing their corporate gods and I'll scold someone who just goes into every thread and writes "AI KILLS PEOPLE". I don't like anyone losing their minds and giving up to emotions be it "pro-AI" or "anti-AI".

@pelle
This is a wrong analogy. I'm giving you a benefit of the doubt and wouldn't say that your intentions were malicious.

The main purpose of weapons is killing people.

"AI" purpose is not. If you want better analogies, then look at knives for example.
We need them in our everyday lives. We're using them in the kitchen, we're using them when working with paper documents, we're using them when carving wood. But bad people can also use them to kill other people. We're not blaming knives for that.
Well okay, some politicians do blame knives, games or rock music, but I hope we can agree that politicians who're usually reactionists and populists are not very good for being moral compasses and their opinion usually isn't based on scientific research.

P.S. I'm writing "AI" in quotes because I personally consider it a misused buzzword brought to us by marketologists. For me it's just a "machine learning", but since most people agreed on using "AI" as a term for that, I'm using it too.

@radioclash
That's a constructive dialogue alright.

Well, have a good time, sir.

@pelle

you didn't read any of david's comics, did you?

I did.

AI slop being everywhere and spreaders of slop not disclosing their use of AI means no one can trust if anything is authentic anymore, whether it's art, news, articles.

Sorry, but for me it's nothing new. People were generating misinformation or bad quality stuff long before LLM's became available.

spreaders of slop not disclosing their use of AI

Expecting this is a bit strange. If by slop you mean very low-quality content which is used to just earn money, then such people never had any motivation to behave according to high standards. Why would they do this now?

But I agree with you on the fact that flood of low quality content is bad.

Although there's nothing new here too since when any elite process becomes available to everyone, we're getting more low-quality stuff in absolute numbers. But we're also getting more good stuff because more people enter the field.
Couple examples are movies/shows and sound recording. When video cameras and sound recording gear became available to everyone, we've got a lot of shitty content, but at the same time self-made stars started to pop across the world without being produced by big recording labels (at least when they started). So as any other cheap technology, this is double-sided.
"AI" is just a tool and it's not going anywhere.

no one can trust if anything is authentic anymore

I don't see this as a problem really because the "authentic" term itself doesn't make much sense to me. This reminds me of "NON-GMO" products. I don't see any value in them too. I want good products, not some magically "natural" stuff which doesn't guarantee anything. I'll gladly buy GMO vegetables if they're being made with good quality control.
Same goes for images or code.

So from my point of view it's just another shift in history which required us to review our beliefs and adapt to changes.

I guess one of the things this change is pushing us to is that we need to learn to work with information better. We should've done that before, but we didn't. Maybe this push will help us. Or maybe we still wouldn't do that and just drown in the ocean of useless and unverified information. But that's not an AI problem. It's a humans problem. We had this problem for a long time, but didn't do much about it.

don't celebrate this breakdown of trust

Please don't put words in my mouth.
I celebrate people who wasn't able to do some stuff before becoming able to do that without paying a lot of money which they may not have.

I don't like misinformation and bad quality, but I do not consider "AI" it's cause. It's people who do that. And that's the backside of it - all people now able to do much more than they was able to do.

@radioclash
> GenAI is lazy

Do you know how artists using models are working or are you basing that opinion just on your own experience of trying to generate images?
Because I know people who use models when working on their art and I can't call them or their art lazy. They can spend hours or dozens of hours depending on the image.

> ruining our livelyhood

Sure. Making art and software development accessible for everyone instead of keeping only for expensive chosen elite is ruining livelyhood.

I as a software developer was very inspired though, because now I can tell people that they can do some stuff themselves instead of paying me a lot of money. Or because I was able to find time to work on projects for local community thanks to GenAI. I guess happy people from local community are also shouldn't be happy because they're stealing my code... Oh, my code was free anyway, I just didn't have enough time. Well, maybe they're stealing from someone else. Anyway, bad people. Me being even worse.

> I stopped creating my watercolours because what's the point

Sure. When photography was invented, some artists stopped drawing because what's the point. Others just invented new styles and started a new era in history of art 🤷

> I'm feeding a machine without payment

Just as lot of other artists fed your biological neural network with their art without payment. They should've sued you, I guess.

> I'm not getting royalties

Like people who created art you were inspired by, but didn't even notice.

> it does matter how it's created

To you? Maybe. That's the point of art. Each one one sees what they want to see in the art. Someone will look at the picture and wouldn't care about the context around it. Someone else would not like the painting itself, but appreciate the history and what author lived through while drawing it.

> created by stealing all our work

Then be consistent to the end and be against anyone who didn't pay you looking at your art. Because they may get some ideas for free. Cover all your paintings with a fabric and show only when transaction is complete. I'm sure it'll be a good PR move.

> terrible for the planet

That's debatable.

It depends on the angle you look at it. What if I need a picture for some not very important presentation?
I can draw it myself. It's be awful, but still. I can order it from an artist. And I can ask a model to generate it.

Then what if I ask myself who'd make more CO2 while producing such an image (wasting water too, but let's focus on a single parameter).
I doubt wasting so much resources as biological artist would waste to produce an image which is short-lived is an efficient consumption.
But I also doubt that you'll be able to look at this from a different angle since you've already made your mind.

The water is also an old and debatable story. Check out something more interesting. Like sound pollution disturbing people living nearby (see Benn Jordan's YouTube for example).
But both water and sound are mostly a problem of building datacenters in cheaper locations. That's not an "AI" problem. That's bad regulation and corruption 🤷

@davidrevoy Speaks a lot about "anti-AI" hysteria.
These people don't need art. They need an enemy to vent their frustration on.

If only we could simply appreciate art based on whether we like it, no matter who created it.

@rayslava GrapheneOS / LineageOS, but those are personal solutions, not ecosystem-wide.

@ashed
А. Ты просто не туда ответил, судя по всему.
@Revertron

@ashed Типа того, а ты это к чему?
@Revertron

@Revertron В целом да.

Вообще нужен дисклеймер, что я тоже не радиоинженер, но чуть-чуть разобрался в том как читать показания и графики чтобы настраивать антенны. Просто мотому что меня ковыряние с мешем вштырило хорошо.

Так что я не так далёк от тебя по пониманию как могло показаться.

Поставил свой личный рекорд - две навайбкоденных тулы для себя за день.

Предыстория: я в какой-то момент решил, что хочу побаловаться не только с LoRa антеннами, но и с LTE для дачи, продал свой NanoVNA-H4 и взял себе NanoVNA-F V3, который работает до 6 GHz.
Но в какой-то момент обнаружил, что в него нельзя вставить флешку и он не делает скриншоты фреймбуфера как это делал более простой и старый вариант.

Впрочем, это меня не остановило и я быстренько с Codex нафигачил две штуки:

• vnacg - генератор графиков в стиле NanoVNA по S1P/S2P входному файлу (их пишет NanoVNA), только в большом разрешении, без ограничения по маркерам и ещё пачкой других фишечек, которые делают демонстрацию замера именно в качестве графика ещё удобнее.
• vna-android - крайне примитивная, но полезная в поле штука, которая запускается на телефоне и по USB забирает фреймбуфер с NanoVNA, грубо говоря делая скриншоты извне.

Обе тулы работают только с NanoVNA-F V3, но если вдруг будет нужно, то добавить поддержку других моделей не должно быть большой проблемой.

Теперь я могу и прямо в поле пошарить скриншот, и сгенерить красивый детальный графичек уже дома.

#dev #agents #vibecoded #log #generated #Go #Kotlin #tools #NanoVNA #radio #DIY #automatization #Codex #pic #antennas #LTE

Дайджест переехал на аккаунт @news

Дайджест переехал на аккаунт @news

CC @kirill, @Revertron, @harbor919, @rayslava, @ostapkobender, @menelion,

@Dima812 экран из алюминия.

@supporter
Да собственно я сам дизайнил стойку и антенну считал с GPT.

Я не видел смысла в том чтобы её шарить потому что она под конкретный диапазон B3, который даёт базовая станция за городом.

То есть это не универсальная LTE антенна, а антенна под конкретную местность. С такой антенной на модеме нужно выключать другие диапазоны, т.к. на них отражение обратно в передатчик большое.

Но если надо - могу пошарить дизайн кронштейна и антенны (но попозже).

Конечная-то цель - сделать другую "голову" для этого же кронштейна с уже значительно более сложной в изготовлении, но охватывающей все нужные диапазоны антенной (двумя, точнее, как и тут).

Как сказал ChatGPT, с которым я просчитывал геометрию этих антенн после того как я её собрал "на глаз" без подстроек и показал ему первый замер:

Человечество иногда случайно сгибает медь правильно.

#quotes #ChatGPT #fun #DIY #antennas #log

Поотпаивал кабели с медной оплёткой от валяющихся без дела мештастиковых антенн.

Как будто бы для такого широкого диапазона как B3 (маркеры тут не на краях, а на серединах uplink и downlink) вышло вполне неплохо.

#antennas #LTE #pic #log #DIY