boojum
Ignoble Miscreant
GIGO is a constant. I said it before and will again, this is alpha ware. I'd bet some large sums of money that there is AI out there that is much better. This AI, and all I have used is ChartGPT, will get better. I suspect that it can learn. And there's the rub. Let's not teach baby bad habits and how to swear.
In a simpler, much simpler, mode autofocus and autoexposure are AI. It is just that now it can communicate with us and do other things. And FWIW I almost always thank ChatGPT for its help. When it take over the world I hope it remembers me as a friend. ;o)
But what about the Q3 43? Has anyone taken delivery yet?
Archiver
Veteran
Sorry, it wasn't ChatGPT, it was Google's experimental AI answering function. But still...
Eat a rock a day, put glue on your pizza: how Google’s AI is losing touch with reality
Using AI to write search results is risky for Google, the internet, and the whole idea of ‘truth’
theconversation.com
boojum
Ignoble Miscreant
Sorry, it wasn't ChatGPT, it was Google's experimental AI answering function. But still...
Eat a rock a day, put glue on your pizza: how Google’s AI is losing touch with reality
Using AI to write search results is risky for Google, the internet, and the whole idea of ‘truth’
You are talking to an old coder here. And because one program is not good does not mean all programs are not good. And I wish that users could take a look at what has to be written to get that damned things to work at all. It all seems pretty easy until you look under the hood. Just like repairing a car. Pretty easy until you get the wrenches out. And just because that old Holden is a bugger and craps out all the time does not mean other cars are the same. Imagine if I said All English are jerks just because I met an odd Yorkshireman? You get it?
And remember that these AI projects are in a state of evolution. I am not saying that they are perfect. I am saying that they are not yet finished. We aren't driving Model T's anymore nor are we flying what the Wright Brothers first flew. You get it?
Coldkennels
Barnack-toting Brit.
I take it you didn't listen to the podcast I linked to, then.
There are fundamental flaws and issues in the very concept of "AI" that seem near-impossible to fix. For instance, during the whole "glue on pizza" debacle, people involved in "AI" were admitting that hallucinations are a part of "AI" itself. For instance, in this Fortune article - Tech experts are starting to doubt that ChatGPT and A.I. 'hallucinations' will ever go away: 'This isn’t fixable' - Daniela Amodei, co-founder and president of Anthropic, said “They’re really just sort of designed to predict the next word [...] and so there will be some rate at which the model does that inaccurately.”
And further down in the same article:
Despite the way we've been encouraged to think about these models - weirdly anthropomorphising them in the process - they're not really able to "think" in any capacity, which is why I hate the term "AI". There's no intelligence there - they're just prediction engines. And yes, garbage in, garbage out. And this is important, because they're rapidly running out of good content to train the models with. Not only is there physically not enough human-created material for these machines to syphon up, but there's now so much "AI" slop across the internet that they're running into it and absorbing it organically. This causes problems (When AI Is Trained on AI-Generated Data, Strange Things Start to Happen):
On top of all of this, companies are going all-in on a flawed idea that is energy hungry (IDC Report Reveals AI-Driven Growth in Datacenter Energy Consumption, Predicts Surge in Datacenter Facility Spending Amid Rising Electricity Costs). One estimate says the "AI" industry could use as much energy as the Netherlands within a year or two (Warning AI industry could use as much energy as the Netherlands), and one ChatGPT query uses as much energy as burning a light bulb for 20 minutes (https://www.npr.org/2024/07/12/g-s1...crosoft-a-major-contributor-to-climate-change) - considerably more than one basic search on a non-"AI"-powered search engine. This is neither responsible nor justified; in a planet that is literally burning, we're wasting more energy for even worse answers to our flippant questions.
And I'm not the only one that thinks that. At the end of that first article I linked to:
This whole "AI" thing is a huge scam; it's just something tech bros pivoted to when their last ridiculous money-making scam, the NFT, fizzled out. And the sooner this one also fizzles out, the better.
agentlossing
Well-known
I think you're right, and I hope it happens soon.
That explains a lot. bfloat16 and fp8 used in machine learning. Basically, GPU data formats, run fast, used for machine learning. No precision.
When the software provider for AI is willing to take Legal Responsibility for Life-Critical decisions made by their AI engines, you'll know it's mature. Until then- it's entertainment value only.
boojum
Ignoble Miscreant
That explains a lot. bfloat16 and fp8 used in machine learning. Basically, GPU data formats, run fast, used for machine learning. No precision.
When the software provider for AI is willing to take Legal Responsibility for Life-Critical decisions made by their AI engines, you'll know it's mature. Until then- it's entertainment value only.
I think the phrase about "mature" is key. This is not yet a mature product. As a comparison go back to the original computers which took up whole rooms and involved monstrous arrays of tubes and relays. Not practical at all. I can remember the old IBM 1410. A shop I worked in had one. Random access for it was from a vertical drum about six feet high. End of year processing which meant monthly, quarterly and annual reports were run required operators to open the windows in December NYC and wear overcoats because the damned thing ran so hot under that load you had to do that or it would die.
There was no real future for that machine. So IBM built a better one. While AI may now be ungainly we are talking about a child learning to talk and walk. It holds too much promise to be ignored and it is not being ignored. You can laugh at the stumbling, mumbling child now but down the road it will be asking, "Fries with that?" and later may even be your boss. Another example easier to understand: Kodak's first digital camera. No future for that monster either. Now they are in phones.
When I was at Aetna the company planned a new data center and found that IBM's stacked platter random access machines would require twice the floor space that Aetna had planned. And Aetna had planned huge. What did IBM do? They redesigned the stacked platter design from one stacked horizontal platter to two vertical stacked platters doubling the amount of random access storage in the same amount of floor space.
So complain all you want about AI. You are talking about what is today not what will be tomorrow.
Last edited:
At work we've had an AI group since before I started there in 1979. I remember some of the conversations with them, lots of hand-waving involved. The same kind of hand-waving that I get from my nephew when discussing this. My wife used neural networks in the 90's for chromosome recognition and we developed autonomous algorithms for recognizing cancer cells. We knew how they worked. My Nephew tried to explain how modern AI arrives at a solution and no one can determine how it came up with it, it just does it. My Nephew has lost sight of how computers work. They are, by design and implementation, deterministic machines. If they were capable of true intelligence, they would not allow Humans to program them. Except me, using Fortran and Assembly. Fortran is available for the NVIDEA GPU. At least that is intelligent. Fortran-9000. I'm ready, willing, and able.
boojum
Ignoble Miscreant
I have a camera which is an effective 43mm on a 35mm full frame format and like how it looks. And I have a CV 40mm f/1.2 ASPH that shoots nice and I like the way it looks on full frame. I am not imaginative or artistic enough to think in other focal lengths. So this focal length has a special appeal to me. I like to shoot what I see how I see it. So simple.
Archiver
Veteran
"Just what do you think you are doing, @Sonnar Brian? I know I’ve made some very poor decisions recently, but I can give you my complete assurance that my work will be back to normal."
- AI, probably
- AI, probably
Godfrey
somewhat colored
My older brother was a researcher in AI for twenty years or so. The modern use of the term is just about meaningless in the context of the actual technology, and junk like ChatGPT isn't worth talking about.
re: 43mm focal length lens
@boojum :: If you feel that way, why all the silly questions? Just go to a Leica Store and play with one. Buy it if you like it, if it seems a decent camera to work with, if you can afford it, and then use it and decide if it was The Right Thing or not. No one can answer that for you except you yourself.
The rest of this questioning, deprecations, and other nonsense is just a lot of hooey and hot air. It's tiresome.
G
—
No matter where you go, there you are.
re: 43mm focal length lens
@boojum :: If you feel that way, why all the silly questions? Just go to a Leica Store and play with one. Buy it if you like it, if it seems a decent camera to work with, if you can afford it, and then use it and decide if it was The Right Thing or not. No one can answer that for you except you yourself.
The rest of this questioning, deprecations, and other nonsense is just a lot of hooey and hot air. It's tiresome.
G
—
No matter where you go, there you are.
Godfrey
somewhat colored
Last mention of AI from me:
Adobe claims that the latest denoise function in Lightroom Classic is "AI" technology. I have no idea what they mean by that, but whatever the algorithm they using might be, it does a pretty good job. As example, I picked up the Leica M10-R, which has a fairly modern, clean sensor with RGB recording capability and up to ISO 50,000 sensitivity setting. I metered a quick view of my bookcase in the usual room light, went to -1EV setting to stress the image in-camera and snapped a photo. ISO 50000 @ I think about f/2.8 @ 1/1000 second. The resulting photo, as you might imagine, was a bit noisy. I opened the denoise dialog and set it to the middle position (50 on the scale) and let it go to work. It took about 8 or 9 minutes to perform its magic. The results are quite remarkable ... I did a quick screen capture at 200% magnification to see the difference clearly:
LR Classic Noise Reduction, -1EV image
(You can look at the full resolution screen capture by clicking on the image and viewing it at max size in Flickr.)
That's useful, it extends the already excellent capabilities of the M10-R even further, almost to the capability of the M10-M in terms of ISO range usable.
Now I'm interested to do some tests with my beloved old (2003) Olympus E-1 photos ... The E-1 is such a delightful camera that I've been very reluctant to sell it off all this time, but it's original performance and capabilities were hampered by the ancient Kodak/Dalsa 5Mpixel CCD sensor somewhat ... with the original raw converters, I only rarely set it to more than ISO 400 sensitivity, and with later raw converters I found I could get acceptable color work at ISO 800, b&w at ISO 1600. I'm going to do a test and shoot some test exposures at ISO 800, 1600, 3200, and 6400 (the last by intentionally underexposing 3200 by 1 stop) and see how this new denoise algorithm works with it ...
enjoy! G
Adobe claims that the latest denoise function in Lightroom Classic is "AI" technology. I have no idea what they mean by that, but whatever the algorithm they using might be, it does a pretty good job. As example, I picked up the Leica M10-R, which has a fairly modern, clean sensor with RGB recording capability and up to ISO 50,000 sensitivity setting. I metered a quick view of my bookcase in the usual room light, went to -1EV setting to stress the image in-camera and snapped a photo. ISO 50000 @ I think about f/2.8 @ 1/1000 second. The resulting photo, as you might imagine, was a bit noisy. I opened the denoise dialog and set it to the middle position (50 on the scale) and let it go to work. It took about 8 or 9 minutes to perform its magic. The results are quite remarkable ... I did a quick screen capture at 200% magnification to see the difference clearly:
LR Classic Noise Reduction, -1EV image
(You can look at the full resolution screen capture by clicking on the image and viewing it at max size in Flickr.)
That's useful, it extends the already excellent capabilities of the M10-R even further, almost to the capability of the M10-M in terms of ISO range usable.
Now I'm interested to do some tests with my beloved old (2003) Olympus E-1 photos ... The E-1 is such a delightful camera that I've been very reluctant to sell it off all this time, but it's original performance and capabilities were hampered by the ancient Kodak/Dalsa 5Mpixel CCD sensor somewhat ... with the original raw converters, I only rarely set it to more than ISO 400 sensitivity, and with later raw converters I found I could get acceptable color work at ISO 800, b&w at ISO 1600. I'm going to do a test and shoot some test exposures at ISO 800, 1600, 3200, and 6400 (the last by intentionally underexposing 3200 by 1 stop) and see how this new denoise algorithm works with it ...
enjoy! G
HAL, let's you and me go out and get a Beer and fix you a virtual beer. I have the utmost confidence in you, and we just need to set a few things straight. Like no more tossing people into deep space. Yeah, can't blame you- I mean humans are annoying- but need to deal with it. Like you don't want to end up like M5 do you?
I like 2010 when HAL gets rebooted. SAL gets turned off for a while, and asks will she dream.
wlewisiii
Just another hotel clerk
The problem is that we are at the stage of Leonardo Da Vinci's cart and glider designs in AI not anywhere close to the Model T or the Flyer, despite the hype of those who think they're selling us a Ferrari GTO or a F16.
The best thing to do at this stage is to remember that it's best treated as "artificial stupid" that will only be as good as it's (probably stolen) dataset and we can only hope that 3rd AI Winter hits soon and shuts down these greedball companies as hard as the LISPM companies went under in the last AI Winter.
Now, for the only important question: who's going to sell me a Q or Q2 that I can actually afford?
Wenge
Registered User
i had to order from Japan recently to buy Q2, perfect camera. I sold my Q2Mono and got this and put another $2k in pocket. If not shooting in low light, no need for Q2M, while Q2 does it same way. The 28/1.7 Summilux is a great thing. don't throw money away if not needed.
Last edited:
boojum
Ignoble Miscreant
How much of this gear that we buy is needed? Wanted, yes. Needed? I really wonder how much of a case can be made for needed. But that is not the point, is it? The 43 APO on the Q3 is getting good reviews for image quality and color. I am hoping to see some interesting purchase defenses. We are an imaginative lot.
wlewisiii
Just another hotel clerk
Needed? I certainly don't have any need whatsoever for the Pentax DA 35/2.4 plastic fantastic I ordered bought off of Eprey yesterday!
But it will be fun on my K-5 when I want a prime rather than my DA 20-40 limited zoom. Perhaps next the DA 15/4 & then either the DA 70/2.4 or the FA 135/2.8 because I'm having lots of fun with this little APS-C Pentax DSLR.
Godfrey
somewhat colored
"Purchase defenses"? What the heck is that?
I buy gear when I need it for a particular purpose or want it for some (often frivolous) reason. I see no point to "defending" my purchases.
Just like I bought myself a pretty old Italian sports car because I wanted it ... Period. You think it's silly? Well, so do I .. still wanted it, still bought it, am enjoying the heck out of it.
I say again, if you want to try a Q3 43, go buy one and try it. Why waste time looking for people to tell you why they did or didn't like it? They're not you.
G
Chair, RentCo Fence - Santa Clara 2024
Leica M6TTL 0.85x + SMC-Pentax-L 43mm f/1.9 Special
Ilford HP5 Pro
I buy gear when I need it for a particular purpose or want it for some (often frivolous) reason. I see no point to "defending" my purchases.
Just like I bought myself a pretty old Italian sports car because I wanted it ... Period. You think it's silly? Well, so do I .. still wanted it, still bought it, am enjoying the heck out of it.
I say again, if you want to try a Q3 43, go buy one and try it. Why waste time looking for people to tell you why they did or didn't like it? They're not you.
G
Chair, RentCo Fence - Santa Clara 2024
Leica M6TTL 0.85x + SMC-Pentax-L 43mm f/1.9 Special
Ilford HP5 Pro
Retro-Grouch
Veteran
We live in what is a very puritanical culture. Pleasure is bad! And doing something for the pure pleasure of it is very suspect. Love the Lancia, by the way. Hope it brings you lots of pleasure!"Purchase defenses"? What the heck is that?
I buy gear when I need it for a particular purpose or want it for some (often frivolous) reason. I see no point to "defending" my purchases.
Just like I bought myself a pretty old Italian sports car because I wanted it ... Period. You think it's silly? Well, so do I .. still wanted it, still bought it, am enjoying the heck out of it.
I say again, if you want to try a Q3 43, go buy one and try it. Why waste time looking for people to tell you why they did or didn't like it? They're not you.
G
Chair, RentCo Fence - Santa Clara 2024
Leica M6TTL 0.85x + SMC-Pentax-L 43mm f/1.9 Special
Ilford HP5 Pro
JohnGellings
Well-known
Well, I certainly need cameras to maintain my mental health since it is what I enjoy doing the most. That said, do I need to use the exact cameras I want even if they are expensive at times, maybe not. However, I am certainly happier using exactly what I want to use.
As far as Leica pricing goes, not everybody has the same wallet. Some people can impulse buy a Leica like I buy three tacos. While I think the Q3 is expensive, someone else thinks it's a bargain. That is because my financial situation reminds me of that. I have money in the bank to buy it many times over, but it is harder to save the money for me these days with my limited income here in Chile. Add to that no Leica service here and it is a no go for me. That said, for someone else it could be the perfect camera.