8by3

I've got a tangent: I'm done using closed AI.

My last role was HoE after a career as an IC, and honestly, using agents is 95% the same as being an HoE or team lead. Steering the ship, ensuring technical excellence, sometimes fixing dumb bugs, sometimes learning new ways of doing things.

But the whole AI business, especially the closed sort is going to properly screw us all if we let it. At least with open models there is a chance of all of human knowledge being a general good. Not just making the ultra wealthy and powerful, even more so.

They'll go hungry, or start a revolution to take back control.

Not to mention that beef/dairy production is a primary source of methane which is 80x more potent than Co2, but constantly under reported due to its shorter half life. Yes in the long run that matters, but its all the more reason that we should reduce animal agriculture right away as the payback is much faster than limiting Co2.

Doom scrolling is part of the free labour we are providing, in exchange for digital goods.

I actually linked a whole article about this point. Yes BP popularised personal carbon footprint to deflect responsibility. But actually in yge short term dramatically reducing methane is one of the most impactful things we can do as while it has a short half life its 80x more impactful than carbon. And the majority of methane comes from agriculture.

We are no longer at the point of do x instead of y. We need to do it all, put fossil fuel comapnies out of business and change your diet.

Totally, a huge swathe of things are better and cheaper when your make them yourself, you just need the time. Shame we've given up all our free time to work so we can buy worse alternatives.

Also "Access to information" is a funny one, most of the useful information is crowd sourced via direct sources like wikipedia and now indirect sources like LLM's regurgitating all human knowledge. Everything a company is selling you is getting worse, one can only assume the same will happen with AI, as soon as possible otherwise the AI companies won't be able to payback on the investment.

I suspect its popular in part, because it feels like fixing "the problem" without actually having to change your own behaviour at all. Don't need to change my gas guzzler, or fly less, or change your diet. I support the goal, but worry its a distraction.

https://tuhat.net/@sbr/p/change-for-the-climate

Maybe "Code was never the hardest part" is more accurate. Code can be hard, though as others have said, 90% of code out there is crud apps plus forms. The other really hard part of getting people to do any complex work together.

Elixir gives you the productivity for humans and machines alike without the performance costs. Plus you get 25+ years of Erlang battle testing as a foundation.

Its not only a question of five eyes access. There is also the concern of being subject to the whims of a regime that might decide you shouldn't have access to services hosted in their country.

Not a fan of the "no AI" signs on things, much rather focus on the positive than what is not included.

The opposite of short exists.

https://tuhat.net/about

Also open source.

Ideally without using substack at all...

Nice, I tried something similar but crowd sourced.

https://tuhat.net/external

Lists chronological posts from everyone that anyone on tuhat adds as an external source. So you can find new writers and the platform can benefit the greater indie web, not just people that use it to post their own content.

The idea started by scratching my own itch as I wanted to read external blogs inside my tuhat /reading view, then I figured why not make it a global discovery feature.

Independent web platforms should want to improve the decentralized nature of the web, not necessarily capture all people for themselves.

Its been trickle up economics for a long time now.

why?

It doesn't have to be _that_ extreme,

Many of us just wanted vaguely what the promise of a good job was supposed to offer. A decent wage, the ability to afford a home one day, support our families etc.

Maybe not everyone but up until about 2022 I had great job security and loved my work. I knew I was extremely lucky. I didn't have a crazy silicon valley salary but I was doing pretty well on any kind of global scale.

Job security went quickly out the window, with it came a sense of insecurity and the dread that most other working people live under. So I'd argue we aren't more sad than most people, just lost the warm bubble of being paid well to sit in a comfy chair and daily get the reward of solving interesting problems.

My my knee jerk reaction is to agree, I really hate the obvious looking AI images, though I suspect ultimately it comes down to having good taste.

As much as I like to be on the anti AI bandwagon, for which there are plenty of good reasons. Some people use the tool (like any tool) well, to compliment their work.

Eg. this author uses great images to accompany their articles (except for one which is obviously AI), its hard to know if they are real or not but they support the text. https://tuhat.net/@jesusrrodriguez/

Regarding businesses, I think the use of bland stock images made most corporate sites awful years ago. AI is just accelerating and making it easier for the rest of us to create the lazy heartless "content" that has been so common from the corporate world.

https://tuhat.net has comments via email, but only when responding to a newsletter post. Then the replies are put in the authors letter inbox connected to the post. Though all comments are private to the original author.

Pre LLM's I used to get this from managers, but what they would do was quote verbatim from stack overflow or some random blog post of how to do a thing with no to little context or understanding of what they were talking about. And it always turned out to be from amongst the top three results from some popular search engine when describing the complex problem the company/team was facing.

Exporting your subscriber list isn't perfect. I did it after moving to https://tuhat.net, and then after posting from tuhat, got emails from subscribers asking who I was and why I was writing to them. The majority of said subscribers, came from an article I wrote about leaving substack, so you'd think they'd have figured it out.

Some of them were far less polite, never mind that it would have taken them less time to click unsubscribe then writing some insulting email to someone they chose to follow.

I'm with you, tech should be a means to an end, not an end in itself. I've been spending a lot of time recently thinking about using tech for good, to do things we couldn't do without it but without being sticky, or gamified, or enshitified.

That's why I built https://kilta.io/. The wolves have spent the last ~25 years splitting people up, convincing them to optimise themselves as personal products, and in so doing undermining real world communities. Kilta has absolutely zero revenue model, its built to be unmontizeable.

But its really hard to build alternatives to the status quo, as you need to break through all the noise, through the slick marketing campaigns, through the millions in venture funding trying to convince you to create "content" for them to make money from. And anything that is trying to heal the hole these products have created actually needs people to use it to give it value.

They've already started requiring you to be signed in to view content, so I'd argue that have already become, not going to become.

So far tuhat's content has been very high quality, the ~1000 word minimum and lack of public comments make its use for spammers not very appealing. And a user only gets one post per day listed on the homepage. There are some more moderation / spam detection tools built in to remove "bad" users.

Additionally for non English users, the language specific pages eg. https://tuhat.net/es see lower traffic.

Ideally European countries would ban all US tech companies and actually invest in home grown solutions. Already some good pro open source policies. Just take it a step further and take a stance on digital sovereignty

Assume because Spain is part of Nato. Though the two aren't mutually exclusive, the Goverment is banning it from parts, not necessarily everywhere.

I agree that's part of the problem, but not the whole picture. The business models behind these companies is what influences the algorithm.

Also there's the designed algorithm that optimizes time on site by decidcing what you see but there's also an implicit algo that is being hacked. Its our basic primate brains that were never setup for the potential of social recognition by more than our village. Checking for a new like, a comment etc, that's what hooks us, then the algo adds the jet fuel on top.

Whats interesting is that this "We don't care about the code, we care about the outcomes" was common management speak for years before AI was on anyone's radar. In some ways it made sense, it was what distinguished your junior programmer from the senior engineer. One worried about the function, the other about how the systems would interact. But it also swept a whole bunch of technical debt and ticking time bombs under the rug. It allowed the ship it quickly and let it be someone elses problem in 2 years when no one can read this garbage, its slow as mud and will require a full rewrite to add one new feature.

AI and its use encouraged use, feels like gasoline poured on that existing mindset.