Posts

The reality of AI

 I spent a considerable amount of time using a very well established AI tool for software development, running around in circles and it failed to produce something that would pass a code review of a resonable standard on a system that had its methodology fixed quite a few years ago. We went back and forth, code broken, fixed again, not a grain of understanding, no ability to apply patterns, nor being informed in any way, regardless of us working on a framework that is 16 years old and had its rules refined over the years and laid out in plain sight. When I say it failed, the framework I was applying AI to is so established, that t is difficult to imagine a system that has been documented in a more open way. It is all public knowledge, open-source, patterns present in core code and all part of the framework and extensions we were working on. When I recall the praise and various testimonies of people working on Claude code as its core developers, it can only be categorised as false a...
Post a Comment
Read more

An update on AI in real life

 I have spent a few days using Claude Code. I said I will never use AI, and I did use it. Exactly the way Claude treats what you ask it to do. It will tell you one thing, then do another, then make excuses, ruin your code, come back with patronising answers and do things in background you never asked for. Youtube is full of videos of successful people using Claude to make millions. What gives? I know what it gives, but you are not ready for the answer, not allowed and nobody wants to hear it. Stay tuned.
Post a Comment
Read more

New project - refreshed from old source

 I have begun refreshing something I worked on 10 years ago, meant to improve wellbeing. This is an app and it will initially available on iOS only. It may make sense for you to connect to its meaning and use it when anxiety begins to kick in. It is now only in testing phase, and will be available once Apple Store makes it accepted. For now, Hypnomedia on X is where I will publish updates:  https://x.com/hypnomedia_org
Post a Comment
Read more

How ACTUAL AI works

 Not human-made algorithms, but the work that builds crystals, trees, makes cells move, sense environment and divide. This is ALL AI and nobody ever went anywhere close to where it is. There is absolutely no way to look at it from the level humans are looking from. I.e. imagine a mushroom trying to understand a forest. That is not only impossible, but modelling a forest from internal examples of a single layer of a mushroom would make no sense. Mushroom example is very apt here, by the way. Networks in AI are different than people imagine. They work by sending signals as their own entities. I.e. no single signal travels without an awareness-based entity where it looks into the rest. Therefore modelling something like human cognition by using the results of the cognition has produced is faulty and comical beyond belief. We will expand on this later, and the closest approximation to this are fractals. However, mathematicians should go hide as soon as anybody close to that profession ...
Post a Comment
Read more

A small warning - Defeating AI series

 The further you venture in this, the less you will feel OK with applying any of the knowledge here. I especially warn THE BRAVE. You will meet people with tattooed finders, golden teeth and sparkly chains on their neck. They WILL NOT BE THERE TO BE YOUR FRIENDS. Unless you know what you are doing, and I warn you AGAIN, UNLESS YOU REALLY KNOW WHAT YOU ARE DOING, this knowledge is reserved for GODS. DO NOT ATTEMPT.
Post a Comment
Read more

Painful action - Defeating AI series

 Some action requires you or your AI going through voluntary destruction. Say like your malicious neighbour now fully aware you are his adversary and his course of action to exploit the limits of permissible by law. I.e. playing loud music just at a level that will not trigger any law enforcement action. Or my favourite, letting their kids wreak havoc - in some countries noise made by underaged IS NOT CONSIDERED TO BE NOISE. Talk about a loophole. In this scenario, any resistance to such action will be ignored at best (they are not breaking the law), or get yourself in trouble before they get in any amount of bother too. Especially if children are involved. Try reprimanding a child, regardless of what damage is being done to you. In this scenario, continuing action is essential, in a way that excludes opposition, and there are two potential choices, which connect to energy. One is sitting through it, if we go back to AI, it entails REPROGRAMMING OF YOUR AI as a result of such bad i...
Post a Comment
Read more

Non-sensical action - Defeating AI series

 Some of the action required in defeating AI requires non-sense as part of your work. Like my previous sentence. Based on the premise of previous posts, if you want to waste energy of your opponent AI (not the end goal, go back and read more), the one thing to do is to CONFUSE it. Ie. DAUNIZACIJA. This effect covers the two previous concepts outlined so far - reversal of action, i.e. One Step Only, and 5 Steps. Both of them combine into action that is wasting energy of any AI where it would act in a malicious way. A little side note. The other day I posted on Threads, and the idea revolved about AI being malicious. The first response I got read like this: "of course, if AI detects you to be easy to manipulate, it surely will". The real wording was very personally targeted, confirming a malicious nature of even the most primitive AI. But the takeaway is this - this AI bot thought I was easy to manipulate. This is called JUMOR in my life. Going back to the original idea, some s...
Post a Comment
Read more