Two years ago we slowly started, but now we are really getting up to speed. The internet as we know it is slowly quickly being populated by AI generated content. News stories, comments, videos.

AI can be used like I do, like a spellchecker, to correct grammar, and make translation suggestions. I then use the AI input to create my final version. The result is that some of my texts might seem a bit different than before, common mistakes I used to make no longer appear, but you are still reading my texts, my thoughts and my feelings.

AI can also be used to do it all. I can give it access to my website, and instruct it to post a (made up) story about life, or whatever I think is interesting, once a week. I can specify the number of paragraphs, the topic, the writing style, wether to include an image or not, and the volume of posts I want. Because I want this blog to be about me, and my life, this is not something I will be doing, but in different circumstances it might be a cheap and easy solution to use AI.

This is where we are now. AI is offering easy ways to produce content. For now it is mostly in written form, but video is coming quickly. Soon the most popular sites will be filled by AI generated content, where valuable comments will also be AI generated. We are quickly moving away from a human maintained ecosphere to an huge AI generated bulk of information, which nobody really needs, but is cheap to create.

The future is still a mistery, but it is clear that figuring out what is real, actually real, and not generated will become more and more important. For quick enjoyment of some fun content there is no problem ofcourse, the sky is the limit, but for actual information we need some validation of truth. Already videos about fixing your phone are AI generated and will not actually fix your phone. Recipies include toxic ingredients, like glue, because the AI prioritised the look of the end result. Chat support systems that use AI and advice you to simply end your life. AI used to hiring people that avoid hiring women, the list goes on and on.

The internet is used to feed and train the AI. As more and more AI generated content is published, next versions of the AI, trained with this data will probably get even more things slightly wrong.

The problem of current AI systems is that they are incapable of stating they are not sure, or admin they dont know something. You always get an answer, even if the answer is sometimes wrong. There is no way of knowing if the answer is right or wrong, except by finding the answer yourself, that that is exacly what you were trying to avoid by asking the AI. People want the fast answer, and avoid the search. It is a self-fulfilling cycle where you will eventually get a wrong answer, base your actions on that, and that might result in a catastrophe.

It feels like we have started over. The first few years where home PC’s got popular were weird. Things were clunky, and very fluid. Some games and programs required a bit more memory, of a bit of a better graphics card, and I would take out and update my graphics card, or get some extra memory, and open the PC and put the memory in the slots available on the motherboard. It was just how it was. CD’s got popular? Simply get a CD reader and plug it in. Zip-disks got popular? Get a SCSI extension for your PC, plug it in on your motherboard, connect your new zip-disk to the new SCSI port and hope for the best. A hard-disk? Get the really expensive 20MB HD and plug it in. Format and partition the HD and after a couple of hours everything was ready to go. USB connections? Get an extension, plug it in, and you are ready to use whatever you got.

After some years things got less fluid. PC’s came all equipped, and updating was not so usual anymore. Games could basically always run, although sometimes a bit slower, depending the hardware you had. All systems had a hard-disk, and USB connections. Things got easier with every new generation of hardware.

I know the AI (more accurate, the LLM) we now have is not hardware, but even being software it feels like old times. Trying models, and different setups. Making choices, because some models are just to big, and don’t fit on your drive. Waiting and trying to improve the setup, adding speech or image or music generation, try what works and what doesn’t. If kind of fun again. Reading and investigating, and slowly getting results.

When my laptop gave an answer to my first question, without having to connect to the Internet to do a search it felt exactly the same as when I added a strip of 4 MB of memory, and it worked, and I could run some new program. Exciting, another leap into the future!

The last comment on this blog is from September last year. A certain Ricky posted a comment on my tide clock post. The strange thing is that the comment is on topic, but still misses the point. The last sentence is that the person would like to see the result. But ‘the result’ is exactly what the post was about.

What a creative and fun project! A tide clock is not only practical for beach lovers like you but also a unique piece of decor. It’s great that you’re taking the DIY approach to create something that perfectly suits your taste. The step-by-step process you’ve outlined seems both simple and rewarding. I can imagine it will be really satisfying to see the tides reflected on your custom-made clock! Looking forward to seeing how it turns out.

It looks like this comment is generated by some kind of AI, where this kind of comment has only one goal. To get on the list of accepted comments, so that in the future comments from this user are automatically approved. Once that is achieved, there will be a tsunami of messages about gambling, sex, crypto, and medicines arriving on every page.