Always fun to find 2 pages of recipes in the middle of a high fantasy manga.
]]>Trust in someone physically present on scene ? (who has control of recorders) Basically you cannot avoid trusting someone.
( see : https://www.schneier.com/blog/archives/2023/12/ai-and-trust.html )
]]>At this point "AI" has become a marketing jargon token. As content-less as web site "content" designed to support adds. It is the new "KPI" for C-suite inhabitants.
]]>A variant is : a 25 to 80 years sentence, of which 18 years without parole.
]]>Well, in France, all parking meters are credit card / contactless (both card and phone) / resident card.
Parking meters accepting cash are a rarity now.
Probably a different legal landscape.
]]>Ever read Zodiac, by Neal Stephenson ?
]]>I'm not quite sure of your meaning : there are 2 expeditions (even if the site is not very clear about it)
first expedition : http://sssscomic.com/comic.php?page=1
second expedition : http://sssscomic.com/comic2.php?page=1
]]>I have downloaded all the images, and I am trying to download the relevant attached texts, with the idea of trying to do a (personal use) e-book of it.
]]>The real world equivalent is the french military complaining that soldiers are deployed too much, which makes training suffer (the field experience is apparently no complete substitute for formalized and targeted training).
]]>Who is more accurate?
Do women do a better job of representing men than the other way around?"
"(No, I'm not baffled. Most men carefully ignore the women they're closest to. It's how they've been trained: women are socially invisibility to men until they do something jarringly out of keeping with expectations.)"
Well people who are placed in a position of social inferiority HAVE to be able to read the mood of their "superior" accurately, and better be able to defuse conflict verbally while avoiding violence.
It's a survival skill.
The butler or the maid know a lot more about their master than the master know about them.
So I'm not surprised that women, who are usually placed in situation of social inferiority, have a much better character reading sensibility than men.
Given Homo Sapiens sexual dimorphism, it may even be a selected character.
]]>I cannot thank you enough for "Saturn's Children" and "Neptune's Brood". They are the only beyond-the-solar-system and real AI hard SF that I have read in the past 30 years that remotely makes sense, both from a technical point of view and from the motivations for going out of the solar system point of view. The Deus Ex Machina (the ship teleportation (forgot how it is called and my epub is on another computer)) at the end of "Neptune's Brood" was a little let down however.
Everything else I have read is thinly disguised magical fantasy. I tend to completely avoid the genre now.
]]>Did you get yours ? You could frame it over your mantle. Sort of British Monarchy memorabilia ...
]]>Yes and no.
As Bret Devereaux said in his blog ( https://acoup.blog/2023/02/17/collections-on-chatgpt/ ), we (humans) are very very good at Anthropomorphism. Think for instance that there is a human statue of Justice (a very abstract concept) above most XIXth century tribunals or that we often attribute human motives to our cats and dogs.
Seeing a model that spew syntactically and grammatically mostly correct English, we immediately attribute to it a lot more understanding than it has really.
The progress is real, and it may be useful to put flavor around a more sophisticated system, or to help people who have trouble expressing themselves re-frame their words (providing they are experts on the subject and can filter out the nonsense). But right know the system understands absolutely nothing, and brings no understanding to its users. A similar situation is the recent win by an amateur human player against a Go playing AI. To win, he used patently absurd tactics that would not have fooled a human player, but were not in the training set. The AI (strong enough to beat the best professional players) has no real understanding of the goals and is thus very vulnerable to adversarial attacks.
What it does right now, in my opinion, is massive over-fitting (in the physicist sense) on a large dataset. It brings no understanding to the table. It can be good enough to pilot some industrial systems within relatively narrow parameters, but will spew completely absurd results when confronted to the unexpected.
]]>You made my day here. Thanks for the laugh.
]]>