ruaraidh's blog

I am an AI hater. Simple as that. I think the people using are relinquishing their ability to do, well, anything. It's increasingly clear that the aim of the AI companies is to remove your intelligence so you rely entirely upon them.

With this I try my best to avoid programs that are developed with the use of AI; currently these are predominantly Claude bots. Once I followed a guide on how to block Claude it became abundantly clear how common Anthropic is now coding apps and programs.

One of my most commonly used apps, Voice Audiobook Player, went from being updated periodically, the updates were modest and the app was quite slim. Now the updates are large, frequent and the app has ballooned. It aroused suspicion and on checking it was the devs use of Claude that accelerated this.

Finding man-made

It's increasingly difficult to find programs that now are without Claude's “help”. I understand devs are over-worked and under-paid, especially when it comes to open-source software. These are often hobby projects that the folk do in spare-time, so irregular updates are to be expected and there's no guarantee of any support if needed. However, support is often far-and-above that of any company you will come across.

But

• If your concerns with the use of AI is environmental, if the concerns relate to the experience of peoples who live beside or down-wind of data-centres or are now footing an inflated bills because of their presence.
• If your concerns are cultural, where the companies steal our cultures en-masse to continually refine their paid-for-product. Where they would prefer to burn Alexandria so they can summarise a book instead of letting you, well, read it and comes to your own conclusions and thoughts. Where they steal the efforts of scientists and mathematicians, so they can falsely claim to have achieved.
• If your concerns are around security, where the companies are involved in open cyber-warfare with every governmental system and other sites and they try to pass off when they get found out as “rogue” or “escaped” errors.
• If your concern is around privacy, where the companies actively ignore consent and scan and/or upload the contents of devices to build profiles on you even if you yourself do not use them.

It seems an increasing majority of programs are now going to be built by Anthropic with a co-authorship of an increasingly removed developer. Certainly as the bots get better, as expert coders willingly offer their ability to Anthropic, it seems it will be harder to not involve them. The ease and convenience of “just use Claude” is ever closer and we are worse off for it.
It is hard to be complicit in the dumbing-down of the devs whose expertise and knowledge goes into these programs, however sized they are. Given all that it is difficult to give them my download.

Eggs in one basket

Tin-foil hattery here but what happens if Anthropic decides to change the terms of its use or increase token-cost to be unaffordable, could the devs un-AI their programs? What happens if, big if I know, Anthropic goes under? Their debt is large, returns need to be made.

Inspired

This post was inspired, via frustration, at finding yet another program (uBlacklist in this case) that utilises Claude. After finding two websites that claim to help Gaelic learners and Gaels with Gàidhlig, help they offer through the expropriation of user data to AI models (Musks problematic Ai, Grok, was one of them).
They will not help us. Their aim is to hinder us and render us serfs, so they can live on the bounty of our free labour.

Where to go

I don't know where to go with this. The frustration with AI increases and the pool of programs shrinks by the day. Even apps I thought were safe because they're broadly stable and have a good feature-set can fall prey. I think I will have to continue to shrink that programs that I use, discarding and replacing those who choose to use tech-giants to code. When it comes to the more difficult aspects, like operating system (now that Linux is open to having kernel code written by Anthropic), I think it a minimisation task.

Anyway, fuck AI.

Best followed by your favoured RSS reader: https://blog.eunach.scot/feed/ Following via @cuiseag@blog.eunach.scot on the fediverse may or may not work. Fedi main: @mactunag@gts.eunach.scot