Not exactly haha. I just want to make the development of individual prompts more accessible and measurable. The real work is stringing these prompts together to make a system like this one.
SYSTEMS
-
Mind as Spark: Building Smarter Machinery and AI Evolution
By
–
I perceive my own mind as a spark surrounded by lots of machinery smarter than itself, but it has built that machinery and used it to build more machinery. AI continues this game beyond my brain
-
AI Model Planning Horizons and Error Correction Infrastructure
By
–
Perhaps they need to adjust the planning horizon of their models, so that error correction is being built into shared infrastructure. In the meantime, you may need to wait a millisecond longer before you believe information from an unknown source
-
ImageNet-12 Dataset Fits in DGX GPU Memory Easily
By
–
yes, imagenet-12 classification set is only 200GB compressed. can literally load it into gpu memory on a dgx box easy.
-
Smaller Models Handling Smaller Tasks in AI Systems
By
–
Yeah that's why the smaller model is doing the smaller tasks
-
OODA Loop: Speed of Decision Cycles Wins
By
–
The basic idea is that, in combat, pilots go through repeated cycles of:
— Ethan Mollick (@emollick) 18 mars 2024
👁️Observation, gathering data
🧠Orientation, analysis of data, drawing on background & mental state
↔️Decision, choice of action to take
🎆Action, making a decision happen
Whoever does the loops faster, wins pic.twitter.com/T8Jfkyh0NyThe basic idea is that, in combat, pilots go through repeated cycles of:
Observation, gathering data
Orientation, analysis of data, drawing on background & mental state
Decision, choice of action to take
Action, making a decision happen
Whoever does the loops faster, wins -
Building Optimized LLM Inference Systems Efficiently
By
–
Learn how to build an optimized LLM inference system from the ground up in our new short course, Efficiently Serving LLMs, built in collaboration with @predibase and taught by @TravisAddair.
— Andrew Ng (@AndrewYNg) 18 mars 2024
Whether you're serving your own LLM or using a model hosting service, this course will… pic.twitter.com/tyCVsi4SKXLearn how to build an optimized LLM inference system from the ground up in our new short course, Efficiently Serving LLMs, built in collaboration with @predibase and taught by @TravisAddair
. Whether you're serving your own LLM or using a model hosting service, this course will -
Time as Observed Rate of Change Relative to Observer Update
By
–
Time is a much more complex thing than a successor relation. It's an observed rate of change, relative to the update rate of the observer
-
Time Perception in Observer Systems and Dynamic States
By
–
Time only exists from an embedded observer perspective, and the observer is likely part of a dynamic that supervenes on evolving states. Hence the observer will perceive a region over which it can establish perceptual coherence (now), a past (memory) and branching expectations.
-
Substrate Operators and Conceptual Foundations of Computational Space
By
–
It makes no sense to apply the concept of time to the basic substrate operators, anymore than to locate the functions that give rise to the dynamics we conceptualize as space them selves in space.