Tim Singleton

[Tim Singleton @TLSingleton1962]

No excuses.

The rockstar who changed music with one arm #classicrock #music #rock https://t.co/Uy43Mj4eFW via @YouTube

[Tim Singleton @TLSingleton1962]

You know, @AMCTheatres, I think if you tried, real hard, with some careful planning and holding your breath like a spelunker compressing his rib cage to climb down a chimney never before touched by human hands, you could probably shoehorn another metric ton of ads on your webpage. Seriously, I really love the dine in theatres because they are awesome. Your webpage? It is more like standing on a corner in New York City trying to get my bearings while street preachers and homeless are screaming into both ears.

[Tim Singleton @TLSingleton1962]

cutest Baby❤️😘 https://t.co/PHynC8PMEg via @YouTube

[Tim Singleton @TLSingleton1962]

@Mishi_051 50

[Tim Singleton @TLSingleton1962]

@sciam Well, that’s the thing about these kinds of problems, isn’t it? Only a handful of folks actually understand the problem and even fewer have the ability to evaluate supposed solutions to the question being asked.

[Tim Singleton @TLSingleton1962]

@Wendy1u17 1980

[Tim Singleton @TLSingleton1962]

@Hoor9234 55

[Tim Singleton @TLSingleton1962]

@hibanoorie 5+5÷5+5=11, if I am following PEMDAS correctly.

[Tim Singleton @TLSingleton1962]

@BrianHJacobson I am MAGA through and through. I have voted for @POTUS every time he has run and would do so again.

Having said that, there is no retort to this other than 'Yep, true statement, accurate criticism.' and laugh because it is both hilarious and true.

[Tim Singleton @TLSingleton1962]

@oelma__ 2

[Tim Singleton @TLSingleton1962]

This post can scare people who don’t know how training actually works.

Collapse is real if you throw out the human data and retrain only on what the last model wrote. Claiming that new human data would not help is a head scratcher. Why would that be so? You add data so the next pass can correct the last one.

The whole point of adding data is that fact finding and truth determination are like taking a limit in an iterative manner.

The danger I worry about more is not the big models turning into gibberish. It is the owner who hands projections to a financial AI without knowing the principles underneath. Models still hallucinate.

That’s a grocery manager given blueprints, a crew, and heavy equipment and told to build a bridge. He has the information, labor, and machines. He does not have the trade.

[Tim Singleton @TLSingleton1962]

@TravelDestiny10 No.

[Tim Singleton @TLSingleton1962]

This post can scare people who don’t know how training actually works.

Collapse is real if you throw out the human data and retrain only on what the last model wrote. Claiming that new human data would not help is a head scratcher. Why would that be so? You add data so the next pass can correct the last one.

The whole point of adding data is that fact finding and truth determination are like taking a limit in an iterative manner.

The danger I worry about more is not the big models turning into gibberish. It is the owner who hands projections to a financial AI without knowing the principles underneath. Models still hallucinate.

That’s a grocery manager given blueprints, a crew, and heavy equipment and told to build a bridge. He has the information, labor, and machines. He does not have the trade.

[Tim Singleton @TLSingleton1962]

@Mishi_051 110

[Tim Singleton @TLSingleton1962]

@tymboyszn 99

[Tim Singleton @TLSingleton1962]

@oelma__ Democratic Socialist

[Tim Singleton @TLSingleton1962]

@chloedreamy 40

[Tim Singleton @TLSingleton1962]

@farman4x 36

[Tim Singleton @TLSingleton1962]

@Anamta_ali_ 0

[Tim Singleton @TLSingleton1962]

@farman4x 6