Meanwhileā¦ having lived through so many new big IT things, distributed computing, Client Server (when the role of the client and server were mysteriously switched 
), SW as a service (which was the reversal of distributed computing 
) etc. etc. Anybody remember the blockchain gold rush of a few years ago? I think John Naughton is dead right about AI.
Though, as an aside, I think Apple innovation actually died with Jobs 
Yes I agree, I have no pretentions to understanding how AI works. I struggle to understand anything.
āit is not driven by bitcoin, as you claimed. Iād like to see your source for that BTW.ā
Look no further than Ancient Mariners post #270 above which is what I was responding to.
āIām not really sure your last sentence isnāt very 1950s, but what the heckā
I never claim to be PC or Woke or whatever the current jargon is. (and I do like the occasional Irish joke)
It *is* hugely power hungry though - the proof-of-work calculations to add a new transaction to the blockchain are repeated multiple times and because only the fastest node gets the bitcoin reward for the calculation people throw ever faster and more power hungry hardware at it. It has been estimated Bitcoin consumes more than 160TWh a year - more than Poland 
Then thereās the fact that BC (in fact I think all cryptocurrencies) are basically just Ponzi schemes, they are worth no more than what investors put in - so if you can find gullible investors to push up the price, then you can cash out. Most āman on the Clapham omnibusā investors in BC wind up losing their stake.
Ha, yes - we went from mainframes with multiple terminals, through to putting individual computers onto desktops back to large compute farms (they are not mainframes so much these days, just very pumped up PCs of one sort or another) somewhere in the ācloudā (i.e just somewhere in the Internet), but in doing so embued the word ācloudā with mysticism so that ācloud computingā has to be held in awe.
A few really big data losses should start the pendulum swinging back again.
Oh thatās definitely true, I just donāt think thereās a lot of it in Ireland.
Whatever tickles your fancy Mik.
BTW, ANTS have finally got around to telling me what documentation I need to add my wife to the carte gris. Because we bought the car after we were married they consider it a change of ownership, so I need to pay for a new CG, but no CT required. A win 
Glad it worked. I struggle with ANTS so now use an agent rather than waste time.
You wonāt believe it. When I went to fill in the new cerfa, I used the info from the original cerfa I used to register the car in 2019 and I discovered I had requested my wife as co-owner, with passport copy and all. Itās ANTS who have screwed up
So, I have said if I have to register a change ownership she should be the primary owner. Iām looking forward to the response
Todays contribution from Gemini!
Unfortunately, thereās no direct conversion factor between MPG (miles per gallon) and km/L (kilometers per liter). This is because the units measure different distances and volumes.
Nah
MPG to km/l is easy enough because both measure distance per unit volume
Multiply mpg by 1.6093 to get km per gallon
Divide that by 4.54609 makes km per litre.
Overall divide mpg by 2.8248 to get km per litre.
However fuel consumption in SI units is generally quoted in l/100km - which is what I think you were alluding to.
To get that take the reciprocal of km/l (or divide 2.8248 by mpg), to get l/km, then multiply by 100.
So, I *think* it all comes out to be 282.48/mpg, certainly do-able though I do need to double check my working out.
BTW if Geminii told you that conversion was impossible, Iād get a better AI.
I was so tempted to reply in the same vein but you beat me with a much more comprehensive explanation than I was going to offer. 
I can do the maths, I was just stunned that Gemini couldnāt.
Wait till they start putting this stuff in weapons or medical procedures etcā¦
The problem is that the conversion probably didnāt appear in training data and, despite the acronym, there is precious little intelligence in modern "AI"s - they cannot make intuitive leaps. In this case not even one that required two steps, much less the reciprocal conversion.
So I think what you are saying is it is the claimed intelligence that is artificial
Non existent more like.
Though it does prove that you donāt need to be intelligent to hold a plausible conversation with someone.
Until Google just scraped it from your post Onwards and upwards.
I was shocked to see Google expect people to pay for this rubbish.
And also how cautious one needs to be when using Ai as a possible source of advice or information.
Allowing AIās to see their own output in their training data makes them worse.
A spiral of doom The ultimate digital echo chamber.
la fin des haricots, je pense queā¦ oui. The end to end bubble has burst.