Tag Archives: DML

The 9mm hard drive

This is a new side to some, the people know one side to any person and at some point that person reveals another side. This is whaat we see (at https://www.ndtv.com/world-news/how-ukraine-war-has-turned-ex-google-ceo-eric-schmidt-into-licensed-arms-dealer-6372469) and the title ‘How Ukraine War Has Turned Ex Google CEO Into “Licensed Arms Dealer”’ now some will all up in arms (to turn a phrase), but the story is a lot more interesting. We are given “Mr Schmidt said that he is now a licensed arms dealer “because of the way the system works”” there is more to this. You see at some point I had the idea to sell the idea of the Chengdu J-20 to Saudi Arabia (for China), it was merely a thought and my ideas are not merely as noble as it might seem. My simple idea was that Saudi Arabia should be able to defend itself from the aggressors (Iran and Houthi forces in Yemen). When America and Europe wanted to halt the defending options for Saudi Arabia. I saw a simple economic option. The defense budget for Saudi Arabia goes into the dozens of billions (all 127 of them)  and me getting a mere 0.1% of that gets me 127 million dollars, simple clean and a nice setting to make really strong friends in the Middle East. This was before the idea I designed, optionally for Kingdom Holding. And lets face it 127 million makes for a nice retirement package. Eric Schmidt has other reasons (he was already rich enough). He and Sebastian Thrun, CEO of Udacity, are making a new venture namely White Stork. The setting we are given is “The idea basically is to do two things- use AI in complicated, powerful ways for these essentially robotic wars and the second one is to lower the cost of robots,” I see an adaptation to the learning (read: Deeper Machine Learning and LLM’s) that Palantir currently has. I think that a union of the two has far reaching possibilities. So what if the Palantir deployed systems are directly updated by drone systems? We are also given “Mr Schmidt reportedly informed that White Stork will mass-produce drones equipped with Artificial Intelligence to identify targets to eliminate the need for ground battles with tanks, artillery and mortar.” I think it goes further (read: presumed) You see, you can set the cost down but the military are more interested in keeping the timeline as short as possible.

Screenshot

You will have seen this, or something like this before. You have three components, the green ones are low in cost, the red ones high in cost. You want them all to be in the red, but the stage is set that you can only have two, the third one should always be in the other field. As people chase to get high quality and fast systems, that solution will always be an expensive item. Armies are not interested in (to some degree) cheap solutions. Not as long as these solutions are fast and high quality. Now White stork is going to seek fast systems and in robotics this will mean integration of information systems, like robotic intelligence systems that can connect to a secure cloud solution, updating the cloud instantaneously by all systems all at the same time. It become (for the lack of a better term) intelligence by wire. Nations will fork over billions to get it and to that degree no one has this. Not the US (DARPA apparently has some developing stage), not Russia and not China. They all have some kind of wannabe status, but they lack a high tech captain of industry like Eric Schmidt. If I can see this correctly within a few years they would all want him White Stork could be worth a whole lot more than anyone ever thought it could be and I think getting this connected to a system like Palantir is close to the only solution out there and the people at the centre of that axial know this. As I see it the biggest bottleneck in the short term will be an evolved non-repudiation system. We can cyber strike as much as we can but that first defence is a non-repudiation system to ward of attacks and that is where Palantir optionally has the system to make it work. Not for one or two systems, but like 200 drones in different campaigns  all at the same time. These systems need more than a simple deeper machine language, it needs LLM learnings and advance machine learning. With cyber systems that cab keep track of it all. This is not a simple solution but a person like Eric Schmidt could keep track of what was needed he might not be alone, but he is the only one in the stage of these arms of technology. 

His wealth might soon equal that of Bill Gates, the arms industry will pay heavily to get this far ahead. Consider that Saudi Arabia increased its military spending by 50 percent to $69 billion in 2023, approximately 23 percent of its total budget. That is to merely get on par with the America, Russia and China. How much do you think these three would pay to get ahead of the other two? The US is requesting $849.8 billion for next year. With White Stork they could easily double that amount. It is that much money that is in the view of some. 

Just my two cents on the matter. Have a great day.

Leave a comment

Filed under Finance, IT, Military

How to measure success?

That is the question is was facing today. It wasn’t about my success (or lack thereof). It was about the olympics. One member (a fellow Australian) was happy because we had two additional gold members over the United Kingdom. But there was something wrong with that train of thought. It was too American. Don’t get me wrong, as I see it it is great to have more golden medals, but in my old fashioned way of life (and thinking) it is weird that the runners up get to live. I must be going soft in my old age.

You see with Australia grasping a 14-12-9 achievement and the United Kingdom holding onto 12-15-19 at present this list could go into any direction. However, this got me thinking. How do you measure success? Don’t get me wrong the gold number are nice, yet it is not a true list of achievement, is it? I have been pondering this and my mind took me to the old 1,2,3 squared allocation. So Bronze counts as 1, Silver as 4 and gold as 9. Now we get to 183 for Australia and 187 for the United Kingdom. UK won by a nose-hair as jockeys tend to say. So is this actually fair? How can medals be universally set? I don’t think that a boxer will accept equal points to an equestrian, in support, the horse will not go along with that either. Still there is a need to give some level of equality especially as the best of the best of the best in any of these disciplines are competing, yet the simple set to look at the golden medals seems wrong (possibly Canadian Summer McIntosh might agree but she just got 3 golden and one silver medal), at 17 she got (as far as I know) a tied second place with a few others all with three golden medals in the French Olympics. 

However I still ponder, is my formula the right one? It seems to be, but it might be my own shortsightedness to think so. 

Still, the question remains, how do you measure success, and not just in sports. In the 90’s I was subject KRA’s (Key Result Areas) and I accepted them as I had no knowledge on how to measure success. Even in customer care and Technical Support these numbers (when applied to the field I was in) made perfect sense. At some point you need to consider what to measure and how to measure it. Medals are a finite point of achievement, customer care is a little bit more fluidic. So how to go about it? The Olympic medallist might have kicked this off, but my brain takes into all directions. So with one movie script under my belt (for assessment with Dubai Media) am I more successful in scripting then all my friends (both of them)? They are not in that field, so how to generalise some metrics? You see we can grab Z-scores but as far as I can see that is a near obsolete approach to matters (perhaps what the people call AI use this) and now we get to the next bit and why I used Summer McIntosh as an example. These were her first Olympics, so how could there be a Z-score of her and how would it be reliable (or relatable)? Previous competitions? These were her first olympics and even in global events the pressures are different. 

And the field becomes even more complex, you see whatever they call these systems based on LLM’s and Deeper Machine Learning, it is either set by a programmer, or set by data and there the problem becomes a lot larger as both are used. Without proper verification and a number of constraints the equation becomes a GIGO rule (Garbage In Garbage Out).

I wonder how much some players consider success. Most will measure success by their ability to bring home the bonus funds. To some extent I accept that, but when you consider how they went about getting that success becomes a larger issue. In this I take the conceptual setting of Awareness versus Engagement in market research. Awareness could be shown how many impressions (or clicks) something gets, whilst engagement requires interaction with the solution. As I have always stated Engagement wins every time, but the large companies often herald views per thousand (or clicks as a secondary). So who get the price turkey at the end? Large Language Models with (Deeper) Machine Learning what some call a version of AI has issues and the world is waking up to Nvidia (not meant in a bad way). You see there is currently no AI, not yet anyway. What there is (the LLM and DML reference) is awesome and it can do great stuff, but it has issues like the legal sector recently saw. There is a lack of verification and that will be an issue in plenty of fields. 

Have a successful day.

Leave a comment

Filed under IT, Media, Science