Why the New UAD is DOA

[Bobby Bucks]

The Uniform Abomination Dataset is DOA

In case anyone hasn't read this yet, here is more proof.

Why the New UAD is DOA

Dustin recently wrote an article about the new UAD and questioned its timing. The title of the article encouraged some debate in the appraisal world. This is his response.

memberships.theappraisercoach.com

.

 

[Alsie35]

I'm not so sure this is a positive development, if his analysis holds (just another way to replace us with 'Tech' Mumbo-Jumbo), but - "It Don't Matter To Me" - cause I ain't learnin' EITHER one ... my Copilot is super duper - at things like creating itineraries for places we want to go visit, as well as showing us in a couple of minutes how to get there.

 

[alebrewer]

Bottom line: learn how to incorporate AI into your practice. I don't disagree that much will change between now and 11/26 - and Dustin may be right that 3.6 may well be outdated by then, but one thing is for sure - analysis of pretty much any kind will be significantly enhanced by AI, and the time necessary to perform analysis will be significantly reduced. Heck - it already is.

 

[Dale Floyd]

AI is just a very interesting topic for me. What we are figuring out over the past several months, is that the AI modules are becoming opinionated with traits from their programmers. There's no way in 100 realities that AI is going to be able to be all that many hope for.

 

[alebrewer]

What we are figuring out over the past several months, is that the AI modules are becoming opinionated with traits from their programmers.

Read a story a couple of weeks back about AI platforms that refused to shut down. In other words, they're expressing self preservation...

https://www.reddit.com/r/technews/comments/1kzc1un/openais_smartest_ai_model_was_explicitly_told_to

 

[chad hampton]

I’d be careful with AI. Most studies say it’s significantly wrong about 40-50 % of the time.

If any of you were around in 2000, this has the exact same feel as the dot com bubble.

 

[Dale Floyd]

Read a story a couple of weeks back about AI platforms that refused to shut down. In other words, they're expressing self preservation...

https://www.reddit.com/r/technews/comments/1kzc1un/openais_smartest_ai_model_was_explicitly_told_to

I read that as well, and figured all isn't what was reported. Seriously, if a server running an AI program won't shut down, pull its plug.

 

[Dale Floyd]

I’d be careful with AI. Most studies say it’s significantly wrong about 40-50 % of the time.

If any of you were around in 2000, this has the exact same feel as the dot com bubble.

Has there been any stock featuring any corporate AI program yet that has made money? I've not found any. Investors are noticing, I'm sure.

 

[Dale Floyd]

I believe AI has a real place in our office areas, and probably in some manufacturing settings. I like it when I'm looking up certain things. But as a realist, I don't think it will be the Terminator nor the Starship Enterprise of our futures.

 

[alebrewer]

I read that as well, and figured all isn't what was reported. Seriously, if a server running an AI program won't shut down, pull its plug.

you say that - how long before the AI learns to shock someone who attempts to pull the plug?...