Try AI prompt: "Can you determine the central concepts of this paper? Are there any statements that appear vacuous and don't contribute to an understanding of the concepts or exposition of the concept." If the AI is baffled, the paper may need help. I've found that baffled AI is frequently a forward indicator for people being baffled. Be sure to ask follow- questions.
One major contribution of Rene Descartes is that he created "Meta #Cognition", the ability to create a #simulation of our own neural net in our mind, with our ability to annotate it as we learned things and pass it to subsequent generations: the foundations of #math and #science.
I'm really feeling Cartesian Dualism today. My #writer self plus the English language feel like a set of children's blocks while my ghost struggles to form the primitive blocks into something resembling the ghost's conceived idea.
This is why I am going to get a guardrail free local model running. This book does not contain any explicit sexual content. #AI guardrails are the nemesis of #LiteraryFiction#writers. They will also flag teens in violent situations.
For example, if you use Google Notebook, despite the advertised 1M token context, I have actually measured it and you are only allowed about 10K. On Gemini, you notice that suddenly if forgets things and that's because the array is trying to scavenge memory because resources are short.
@TheZachMueller Because there are a lot of games played with cloud AI out of sight that effect the accuracy of your results. AI is not just for coding. If you use AI for something other than coding, you hit an accuracy/hallucination wall rather quickly. And it's due to resource throttling.
@TheAhmadOsman@theo Some of us are stuck back on earlier dense models because new models seem all about coding. But if you are using #AI to analyze literature, it takes "nodes of knowledge" to catch subtleties in the language. MoE actually limits the practical usefulness to "logical" tasks.
@nytimes Moving the state of corporate residence is easy to do. They are already looking at real estate in Tennessee. The meat on the corporate bones can follow later. It doesn't have to occur suddenly like a switch. Tesla moved to Texas but they still have sites in California.
#AI is a nondeterministic program. The closed AI companies release stories of doom to distract from the fact that when their program damages things, they are liable and can be sued. Their programs damage things due to pure engineering incompetence. That's the beginning and end of it.
I am so tired of #Microsoft#Windows11 update hijacking my machine for literally hours. I have lost 2 hours to this update so far. I'm really looking forward to getting on my new Linux workflow. Productivity without bullshit.
Be aware that with the interest rate hikes happening that the most fragile of AI debt could start to buckle. What I would expect as a first sign is a curtailment of capital equipment purchasing. This could stop the hardware price hikes by the spring. This is an unstable situation that bears watching. Hardware price arbitrage is one of the most dangerous games to play with money.
Be aware that with the interest rate hikes happening that the most fragile of AI debt could start to buckle. What I would expect as a first sign is a curtailment of capital equipment purchasing. This could stop the hardware price hikes by the spring. This is an unstable situation that bears watching. Hardware price arbitrage is one of the most dangerous games to play with money.
#AI#debt is fueling #inflation. Purchasing things faster than they can be built is very inflationary. As a result, the #Fed#raised rates a quarter of a point today and penciled in another for this year. Rising interest rates eventually crashes a debt #bubble. Here we go.
@jun_song Be aware that many companies, including Amazon, use the same supplier that is a known scalper. Look for industrial suppliers and expect to wait at least a month before the card ships. If you can get it soon, it probably means it's supplied by a scalper.