Posts by Locomotion69 236 publicly visible posts • joined 28 Aug 2014
Autonamous cars work great in a homogenous environment with only other self driving cars, going the same direction. For example, one lane on a motorway.
All other environments are too dynamic to be reasonable safe. I mean, we know how to drive for decades and still people die on the roads, by unfortunate accidents or by misbehavior. And from this data, they teach their AI.
There is actually some logic behind this - from the point of view of Mr. President:
1. Knowledge is dangerous once in the hand of the enemy (especially if it is a personal or political enemy, or can become one)
2. The USA needs minions to populate the returning factories (cheap labour).
Me, sinical?
and since, the world has changed again.
Congratulations Java, you've reached the age at which you are called a "niche".
I did do my share of Java in the past, but never really liked it or found it well suited for the purpose we were using it for. But in those days, it was "fashionable" and "easy to get experienced developers".
If Apple would transfer (significant) volume to be manufactured in the U.S. of A, would it make it a cheaper product for the American Citizen?
Would an imported iPhone / iPad with 25% Donald Tariff be more expensive? I cannot tell.
And how long does it take/how much does it cost to implement local manufacturing and supporting logistics?
All I can see is that the label "made in the U.S. of A." does not necessarily make it a cheaper or better product.
Mr. President is a real estate entrepeneur witth no/limited experience in manufacturing or supply chain management. For once, I believe mr. Cook is right and buying time is the best way out for now. In less then 4 years time there is another Mr(s) President, with different views and dito politics.
AI has is merits in code generation - but to an extent. Knowing its power and therefore its weakness is key in a development project. That is what today's AI specialist should be accounted for.
You can do with less coders in the end - they are swapped with more "system integrators", busy to glue all these AI fragments into one coherent product that ultimately does compile and run. Understanding code remains necessary to be succesfull at the job, and understanding any piece of code these days is often a challenge on its own - with or without AI.
Currently I have a license for an ancient version of Elements & Premiere - suits me for years.
Yet, with the new plan, I can get Elements & Premiere, with a 3 year license, not renewable, nothing. So I pay for software that stops working in three years to come.
Sorry Adobe, but this is b***s**t. I know: Your product, your terms, but my money, so no new Adobe products for me. I keep my current old software until the hardware on which it runs fails.
Thank you & Goodbye
Sure, but someone has to read the paper to recognise that's what it's doing which appears to be one of the current use cases for AI in the first place.
So, we need to apply AI to determine IF some input is trustworthy/reliable/correct/real science/...<your criterion goes here>, THEN we take that to train the LLM.
Let's make a LLM for that analysis...... OK, I get my coat -->
Given several existing standards for software development (I am not talking about coding standards but more generic development standards like CENELEC 50716, IEC 61508-3 or DO-178C), there is more to do than just "generate the code" using AI. You might want to use AI, but the design in front, and the checks afterwards cannot be assigned to AI in its current state and trust.
There are opportunities left, although this kind of software engineering may not appeal to everybody.
The new name is Metareverse then.
And the industry fails to understand that its existence is based on them providing solutions to the needs of the people. And yet, they keep introducing solutions to a problem nobody has.
Neither the metaverse nor AI offer solutions to real-life problems to the extend that their cost outweight the savings.
Now is the perfect time to decide which future will keep you happiest and most productive, a choice as much psychological and technological, and find a way to have fun as you take your chosen path. It's that or going to eBay for that Commodore 64.
Which is actually a nice thought - if happiness and fun are to prevail, a Commodore 64 perfectly fits.
There are a lot of organisations having your data with no means to escape: government (both local and country-wide), insurance companies, health services to name a few.
All of these are "on grid" to some extend - and can "leak information", or be hacked.
The only "good" thing here is that nothing really bad happened to the soldiers out there in the field because of this cock-up.
And what other "mistakes" have been made that we are not aware about (yet) ?
Imagine what it is going to look like if you are out there, serving your country, and be massacred because your upper commanders are too incompetent to give a sh*t about operation security.
(Me glad not being in the US military)
There is, however, an "opportunity" to help consolidate, modernize, and reinvent the federal government to drive a whole new level of "efficiency,"
I hope Mrs. Sweet is not referring to this in the meaning of "making our own services obsolete". Accenture's shareholders will not be happy about that...
If there were a viable economical and commercially attractive alternative to these US services, it would have been available by now, wouldn't it?
The point is that the investment required to achieve the level of flexibility and functionality provided by US companies is way too high to ever become profitable.
Calling for the government to support funding is too easy in this regard. If dependency of US-controlled companies is regarded a risk, you have to act as company.
So if being independent from the US is now important, let them start a cloud/service business themselves as an alternative to the US controlled companies. Do neither count nor wait on the governments for action.
Apart from having FRMCS right or wrong, this is about the assignment of the frequency bands over different applications.
If you want trains crossing borders, they have to comply to the basic parameters - like using the frequency bands for the correct reasons and under correct conditions.
The UK has cross-border rail connections to the EU: the Channel Tunnel Rail Link, and not to forget Newry - Dundalk.
This is interference, and that is not so uncommon as you would believe. Almost all higher-frequency electronic equipment emits and is receptible to some electromagnetic radiation.
EMC conformity can be a pain...
In practice, it makes sense not to put different electronic (transmission) equipment close to eachother, especially when you are not providing decent shielding, hence the "open case".
Email cannot be considered an "app" as such, in fact I believe it is one of the most productivity-killing tools in the world.
About Excel: I noticed quite a lot of people using Excel as a word processor or database system, "because it can do it". But a "killer app"?
So Nadella is looking for the "one-app-fits-all-purposes" app?
I guess they recalled it because the calculation of the remaining charge (all done by CoPilot based on Recall-generated screendumps) did not match the actual value.
Then people were surprised by a sudden shutdown while still "in the green".
No really, how hard can this feature be? It is not that laptops are new to the market, are they?
The value is in owning the data - it is yours, and yours only.
There is an anology in owning huge masses of land. You can use it for farming, housing, explore the soil, do some mining for rare earth materials, whatever you like to see if there is any added value to the data. But you don't have to. And if others want to do something with it, they need your approval (and give you money).
Doing data analysis this way is like determining what to do with the land - giving the risks and opportunities (return on investment) with all options you have, you decide to do a little more research... (repeat).
A billionaire with too much spare time has invented some sort of A.I. that is capable to install itself in the human brain.
As this billionaire is involved in the election campaign of one of the worlds largest nations, he is able to install the A.I. in the brain of a president candidate by means of a murder attempt.
The candidate happens to win the election and comes to power. The A.I. immediately activates and decisions are dicated by the billionaire.
MI 5 detects weird radiation coming out of the president's head, which after examination appears to be an encryption communication channel using 7000+ low earth satellites connecting the president to the villains HQ.
Bond has a double mission: disturb the up- and downlink to the president and trace the villain to destroy the villains empire.
To escape, the villain puts on a spacesuit and launches himself into space, sitting in a convertible car.
I know exactly where AI will be used for: more data harvesting (on you) so that AI customers (not you) can be informed (about you), and they will be able to target (you) with adds and "exclusive offers", a.k.a. going after (your) money.
That is why "everybody" should have an AI-enabled PC - to do your own harvesting on their behalf.
Maybe I am too pessimistic.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
