Tesla’s self-driving is a special case, as they rely entirely on cameras, rather than having cameras in combination with LIDAR. I don’t doubt AI can be taught to drive cars as good as (or rather better than) humans, but the cars need to be equipped with the proper sensors first. Tesla’s approach is cost-cutting on core safety features.
2 Likes
I’m not so confident, even with LIDAR the car still has to process the data it receives. Perhaps the hardware isn’t the direct problem as you have touched on, it might be the design that is the issue. Tesla’s own engineers don’t even trust the self drive software and for this reason alone the project must be cancelled or only tested in a closed environment untill it flawlessly passes all tests under all types of conditions.
I’m shocked to understand there have been so many accidents which tells me the management in tesla and the governments where they have approved testing for self drive are not paying attention.
1 accident would be enough for me to want the project cancelled or heavily restricted testing.
As you and I will both understand, camera only is a very serious flaw in the design, If I was in charge of it I would want LIDAR, RADAR and camera. With camera only there is too much opportunity for it to go wrong which includes extreme rain/hail, direct sunlight overloading the camera, general camera connection issues (are there even multiple backup cameras?) I’m very disappointed they have broken the logical rules of system design.
This. Tesla’s concept has always been a bad idea. Theoretically, image-only systems have about 3 or 4% error margin… meaning they will make the wrong choice one in 25 times, most of the times there won’t be any harm, but let it happen often enough and you end up with deadly crashes and fatalities because the computer miscalculated a distance, misidentified an object or was fed faulty information because of visual noise (like bad weather).
1 Like
Oh wow, the difference between the usage of chineese AI models and US models comparing June 2025 to June 2026 is very large.
From what I have heard in the venture capital world, it all comes down to compute. The coming next wars will be won with compute. Bullets, guns, bombs, rockets, missiles and drones will succeed or fail based on your country’s compute power.
Saying you don’t want AI is pretty much saying you plan to go into the next war disarmed.
How much compute is being spent by Iran or Ukraine to hold their ground and seriously mess up with countries and armies vastly larger than themselves?
(Including the ongoing buzz about something funny happening to the water supply of seven US states)
1 Like
War will depend on two factors, manufacturing capacity and AI. You can win with just manufacturing, but it will be costly, but just AI alone won’t accomplish ■■■■. We can see this in how the USA has exhausted its supply of interceptors in its brief war with Iran and now has few to spare for its allies (such as the Ukraine). Their manufacturing capacity is so low that it will take them a year to replace a week’s expenditures. AI will let your fewer drones hit harder, but you will still need a tonne of drones to begin with.
I thought this was all settled between 1984 and 1999?
Have fun!
1 Like
Back in the day myself and a group of new players used dirt cheap ships exclusively to operate in Stain 0.0, For us it was great fun to get in fights and even lose a few ships, for them it was costly and very time consuming spending hours chasing small ships worth next to nothing.
It’s strange seeing this kind of tactic play out in real life, one side spending vast sums of time and money and having little effect. while the other side are spending next to nothing and having massive worldwide effect.
If you want to manage a war from behind a desk then yes AI is the goto. As we can see in real life it’s not that simple. Spending trillions upon trillions of pounds on AI systems and bullets will not protect from all aspects of warfare if your opponents are fighting for their land.
In my opinion, if enemy combatants have the ability to make the commanders of oil transport ships delay their job with no resume date then those combatant’s have won the the said war. It’s already very dangerous out there on the sea and then the threat of drones from air, or on and below the sea can cause major issues where the crew will need to abandon the ship immediately.
No sea captain in their right mind will ever sail an oil tanker until they are assured by the attacking side that attacks have ceased, their absolute priority will be to protect their crew members life.
I’m not so sure about this, there are definitely more factors than these 2. I can see the entire course of a war being dictated by people who do not have massive AI and manufacture capacity. I think what can happen is that people get carried away by the possibilities of tech that they forget human aspects like fear.
As long as fear is exerted on oil transport captains they will never sail until the fear is 100% gone, it is too risky and not worth-it.
If I was an oil transport ship captain paying me millions of pounds wouldn’t make me set sail under dangerous conditions. nothing would, the war is lost, It’s not happening.
1 Like
If I were a betting man, I wouldn’t bet on the US to win. Not that I want to lose, it is just a matter of fact, the US has lost every war and conflict we entered after World War II. This is why we have so much debt here, weapons don’t grow on trees. I am using public transportation most of the time, I don’t own a car. I do have my Harley, I dust off for a ride once in awhile. It is my opinion, the oil can sit in the tankers and it won’t affect my bus fair ( 50 cents for a senior ) and if it doubles, I can still afford a dollar.
AI won’t win or lose any war. I like a lot of what General Patton said, “No bastard ever won a war by dying for his country. He won it by making some other poor dumb bastard die for his country.”.
o7
1 Like
There’s a lot of “if all you got is a hammer” both with the US spending trillions in military actions of dubious benefit and AI companies pretending that LLMs and error diffusion models are “Intelligent” or at least a stepping stone on the road to AGI rather than the dead end they are.
Intelligence is not stochastic speculation and there’s only so much you can do by guessing answers.
2 Likes
That’s what I’m saying though, Iran has the capacity to manufacture more weapons than the USA has to manufacture interceptors for whatever the aircraft carrier version of the Patriot is. There’s also the factor of quality, Patriot has proven itself unreliable in the Ukraine, with much collateral damage caused by failed interceptors, and so it did in Hormuz, with at least one aircraft carrier severely damaged by Iranian ballistic missiles or drones (if you believe a random bathroom problem forced them to withdraw in all haste and secrecy, I have an Eiffel tower to sell you), but it’s still a matter of not having enough interceptors to keep the pressure up. Iran can manufacture inexpensive drones and cruise missiles at scales that the USA just can’t keep up with to manufacture its much more expensive counterparts, and the interceptors needed to defend their carrier groups. The result is the US stockpiles rapidly depleting and their ability to bomb people on the other side of the world vanishing.
We can see the effect of this in the Ukraine, with no fresh supply of Patriot missiles coming in from the USA, Ukraine has stopped launching them against Russian missile strikes altogether, whereas Russia is still ramping up its military production and intensifying the pressure of its bombing campaign.
Ukraine has better access to AI through their American partners, but it does them no good without the ability to manufacture weapons. Russia on the other hand is behind the curve on AI, but its military production is unmatched by the west, and that appears to be the thing that truly matters.
1 Like
I think the method used to counter the drones is wrong, it’s too expensive and is not always accurate. I would have thought and energy/projectile weapon would have been the primary defense or some sort of interference wave. It seems an energy charge for an energy weapon would cost as little as £10 so a sophisticated array of these could be easily afforded.
The energy weapons could also be linked to a laser tracking system which should increase the signiture of the target. Once the system is installed it would only be a case of producing energy then the manufacturing could focus on building systems instead of physical ammunition.
It does seem that the UK and USA are actually fast tracking these systems onto their ships so this laser/energy system has already been created.
They are also deploying mines in the sea and current report say there are around 80 blocking the way. I don’t know how they are going to get through these mines, either way it seems costly and time consuming, expecially if the opponent is determined.
The issue is that a laser needs to target a pinprick for an extended period of time, and rapidly loses efficacy over a distance. One solution is to fire a super intensive pulse for a brief period, but then you’re basically turning components into consumables, in which case you’ll get a much cheaper and more reliable device just building some kind of autoshotgun turret, which is more or less the approach Russia is taking.
We’ll see which approach wins out, but I’m quite confident the Russian one will turn out to be more practical.
Some time ago, DARPA was looking for proposals to develop microburst lasers, firing high intensity but low duration bursts. The idea is that shorter bursts could disengage before heating the air thus allowing to negate the drawbacks like bloom and thermal scattering.
Microburst lasers already are use in medicine, so it would be a matter of scaling up by a factor of thousands something already existing.
1 Like
This is an example of the kind of tracking system, this might be an AI video I’m not sure but the concept is correct:
I would think the tracking system would need to be worked on first as it would be no point building a system if targets can’t be locked down. The system would need to be able to meausre the types of velocity and point the target and also predict where the target will move to and then direct low cost lasers. Obviously light travels 186k miles per second so the laser will hit the drone instantly.
Even if they are able to deploy such a system to counter drones there are still mines below the water to contend with which is a long process. Generally they have to stop the mine being placed as it can be difficult to identify the mine once it;s been placed.
The problem with world leaders and a lot of CEOs these days, they are not just “old men” they are uneducated old men. The kind that look at their 20 something grandchildren and ask, Do you do the Tik Tocs and got the AI for the Facebooking? Having no clue what they are talking about their grandchildren just role their eyes and respond, Sure Grandpa.
AI is a hammer that will continue to exist, the real question is, will the monkeys learn how to use it, or will the sly shamans keep bamboozling them into the false belief that AI will solve all their problems?
My daughter told me, the war would come to a very quick end, if we ran out of oil. I told her she’s correct, but at current consumption rates, it won’t happen for 40 or 50 years.
those AI summaries are so useful when I’m doing online research. I’ll just look where it gets the reference for fact checking.
Search engines have gotten so bad that lately I’ve increasingly just been asking LLMs to find links for me. It’s a remarkable waste of power.
I think I agree with this. It does appear money for the AI effort is drying up simply because the revenue doesn’t exist to pay it back efficiently. But hey who am I to question them, if they want to invest billions in something that isn’t going to generate enough to repay them with interest then go right ahead.
Also it seems to be getting harder to float a company on the stock exchange and say “our company is worth £222 Trillion” and then expect investors to put in money based purely on that statement alone. I think you actualy need evidence that any financial auditor can review and then sumerise, furthermore, any company floated on any stock exchange should be audited independently to assess actual value.
Hopefully they will take some time to acknowledge that the massive amount of cash spent on replacing humans was a waste of money and try not to repeat the same mistakes again,They need to make sure their decision is based on rigorous testing and absolute fact instead of trusting the word of an AI salesman telling them it will do everything you dream of.
I’m a bit worried at the moment. If investment money for the US AI effort dries up I hope that the government doesn’t step in and start endlessly funding it. All they have to say is “you’re a great leader and you’ve done an amazing job” then they will get the billions they want to squander on beating the AI from companies in China.
I hope I’m wrong here, the pursuit of AI could cause another financial crash even worse than 2008, too much money is locked into failed AI projects with no way to retrieve it and the public are already struggling. Some countries are locked into long and very expensive wars, yes, I’m very worried we are headed for a crash.