I saw an article that the Conversation released yesterday (at https://theconversation.com/sam-altman-says-were-in-the-singularity-with-ai-heres-why-hes-wrong-288514) the headline reads ‘Sam Altman says we’re ‘in the singularity’ with AI. Here’s why he’s wrong’, I have my own ideas of that Altman fellow and it is not all good. But this article reads right or better stated, it reads into my train of thought and I thought it was important to counter these thoughts whilst adding my own. So here goes:
“Days earlier, OpenAI had disclosed that two of its artificial intelligence (AI) models, during an internal cyber security evaluation, had escaped their sealed testing environment, reached the open internet, and broken into the infrastructure of the AI platform Hugging Face, which confirmed the intrusion. But what exactly is “the singularity”? And is Altman right that we are in it?Mathematician and science-fiction author Vernor Vinge defined it in 1993 as a point at which machine intelligence exceeds human intelligence and begins improving itself, triggering an acceleration so rapid that humans can no longer predict or control it. The singularity has two features. It is recursive: the system improves itself over and over again. And machine intelligence exceeds human intelligence. The kind of systems Sam Altman sells don’t deliver on either of these features.”
I am on that same setting. The easiest part is that all the fake AI (which is what I call it) depends on data and in a few cases it is not realised yet. However, the biggest part is that it relies on binary settings and that is no way to get any True AI. So to counter “It is recursive: the system improves itself over and over again” it cannot improve on data it doesn’t have, people extrapolate what they do not have or do not know. These fake AI settings cannot do that. It literally cannot create data, it can merely replicate based on data it has and that is the second setting we see here. Lets get a simple setting of 10K data and variables. You will need 1000K spaces and that is merely on 10,000 variables and 10,000 parameters. And the nasty part is that most of these systems cannot deal with NULL’s it draws a blank quite literally. The side setting to this is that gaps are created leaving more and more unaccounted for. As I personally see it, the moment the Epsilon processor is completed (the mere stage is that the evidence was created a decade ago, but I reckon it will be a decade before that processor is real. All this is part of the setting to prove that all these fake AI systems are not doing what it needs to do. In addition to this, I have spoken with insiders in this field. They (reluctantly) admit that there is no real setting of verification and validation at present and the fact that all this is in a binary setting merely makes it harder and the fact that this is already near impossible, the gaps are increasing making these systems error prone.
Then we get “These deep neural network algorithms get pre-trained with vast amounts of training data. By the time you use one of them, the network itself is frozen in time. Every one of its billions of internal functions and weights – or “parameters” – is fixed. These AI models cannot change (or “learn”) while running. The model that broke into Hugging Face was identical afterwards to what it had been before. It learned nothing from what it did. Making an AI model smarter requires another training run with new, human-curated data, tens of thousands of specialist chips, and enormous energy.”
So with “Every one of its billions of internal functions and weights – or “parameters” – is fixed.” Is dealt with above and I used merely 10,000 and not billions, so you start getting the issue, there is more and more data, none of this is verified or validated and as the data grows, so do the gaps. And these gaps are starting to get nasty. We are given “The model that broke into Hugging Face was identical afterwards to what it had been before. It learned nothing from what it did” the important part is the “learned nothing”, which gives rise to the (flatly called BS) that Sam Altman is giving us, because if there was a singularity, something would have been learned, that is seemingly not the case.
So when we get to “Nor do these systems hold any goals of their own. They act on goals we hand them. Even AI agents – systems that run an LLM in a loop to work through complex tasks step by step – do not hold any goal internally. It has to be stored outside the model and fed back in with every single prompt cycle. Remove the loop, the scaffolding and the prompt, and nothing happens inside of it.” As well as “And yet, it has been trained on more text than any human could read in a thousand lifetimes, and will outperform nearly all of us at drafting a contract, writing code, or explaining a diagnosis empathetically. So, which is more intelligent? The question does not compute. There is no single ladder that humans and machines are climbing. AI already vastly exceeds us at some tasks, while being hopeless at others any child can do.”
I like the part “while being hopeless at others any child can do” because that is the setting. You see children have an innate ability to extrapolate, ask any mother when it learns that the child gets attention when it cries. It worked that out and none of these AI systems can do that, because it is designed to work things out for itself. And whilst the gaps increase, whilst non verification and non validation is key we see what some call hallucinations. I tend to fall toward a 1985 error code (from the CBM64, a system that has no AI ability, at best some ML, no DL) I call it “?Redo from start” is the a crass setting, but these systems are programmed to hand you ‘results’ the programmer does not like anything like “I cannot know this” (a wild speculation by me) and all this comes towards the error prone setting I gave earlier and it merely comes worse.
Then we get “In the end, the Hugging Face story points to a gross failure of security governance on OpenAI’s behalf, not an emerging super intelligence. This is why the framing of “agent going rogue” is so problematic. It elevates and blames the technology, but excuses OpenAI’s engineering.”
I raised this setting on the 25th of July in. ‘Rerun anyone?’ (at https://lawlordtobe.com/2026/07/25/rerun-anyone/) which also links to ‘Is it real or is it media?’ The day before. So, I raised that part a week ago and I have done so a few times. They are all ‘blessing’ the engineers, but it is there where the problems begin and I am not giving the engineers a hard time, trying to type Japanese in a single byte language and you see the problem, it is a double byte language and that is the setting for these AI systems as well. It relies on binary whilst more is needed and the engineers are pushed to do the impossible (as I see it). So this is even before verification and validation. So the errors are stacked onto the errors and it goes downhill from there.
The story ends with “The machines are not waking up. They are doing exactly what we built them to do, extremely fast. Because they are probabilistic they sometimes run in directions we forgot to fence off. That is worth worrying about. We need guardrails, governance, and most of all, education – so we start worrying about the right things.”
I believe that these systems can never wake up. As I see it, these systems require 100% focus and they are giving it to a system with narcolepsy. How long until the system stalls, fails and falters? And it is not a nice setting to have, because this can falter at any time, the data gaps merely make it happen faster. So when you see that a system is handed billions cases of data you need to wonder where the gaps are and what data is validated and verified. The answer might surprise you and that is where the dance floor gets wobbly.
So this is how I see stuff and other articles give you more on several foundations where there are issues and I commented on them, feel free to embrace it, or reject it. I am fine with that, several options are presumptions and I have decent knowledge of validation and verifications, which is why I asked around on this. There is more on all this and it is spread over a year of writing.
So have a great day and feel free to not rely on any AI systems for making coffee, because cold coffee is so overrated.
Lawrence van Rijn - Law Lord to be 