They work great, better than a fan because it clears the air.
I have one blowing gently across my face while I sleep. It cools me down, and blocks out the highway noise.
146 karma · joined August 22, 2021
I have one blowing gently across my face while I sleep. It cools me down, and blocks out the highway noise.
I worked with a proprietary jet engine analysis software suite written in Fortran 77 and It felt very primitive, albeit fast.
Today I feel like C is a much better choice for speed and readability.
If you really need the fastest performance I dont think Fortran gives that much of an advantage writing assembly.
What were seeing is all models failing to ace these tests.
"Benchmark Saturation" is term that promotes lowering the bar.
This is not "Acing" a test, this is hitting a wall.
Most "AI" is really an optimization algorithm in software tools, same as its always been. This really isnt anything new, aside from adding a chatbot / MCP interface to the same tools.
If the Pareto rule is any indicator, 80% of results come from 20% of causes. It seems that we have alot more to learn about intelligence.
I am reminded of a statement on truth from an ancient philosopher, this sentiment seems to be exactly the opposite of the LLM training paradigm
“The seeker after truth is not one who studies the writings of the ancients and, following his natural disposition, puts his trust in them, but rather the one who suspects his faith in them and questions what he gathers from them, the one who submits to argument and demonstration and not the sayings of human beings whose nature is fraught with all kinds of imperfection and deficiency. Thus the duty of the man who investigates the writings of scientists, if learning the truth is his goal, is to make himself an enemy of all that he reads, and, applying his mind to the core and margins of of its content, attack it from every side. he should also suspect himself as he performs his critical examination of it, so that he may avoid falling into either prejudice or leniency.” - ibn al-Haytham
People have owned property, and worked since the dawn of time, the core original idea in the wealth of nations is that competition lowers prices... thats where the idea of markets come from.
The only real "benefit" come from competition through lower prices, the rest of the factors ie property and labor in a free exchange are the requirements to reach the optimal outcome.
1. corporate consolodation - the only promise of capitalism is that competition lowers prices. Competition is at an all time low.
2. economics and the fallacy of unlimited growth. We live on a finite size planet, and every company is acting as though it expects 10% growth to increase forever. If demand increases and supplies dwindle thats one large prevailing force that would mask as inflation.
The establishment and universities are failing us as they refuse to critically study and address these major conceptual issues.
I'm not even going to touch your examples. They are completely illogical? Dell? The Internet?? Thats one company, and a technology.
If you read the wealth of nations, the only real promise of capitalism is that competition lowers prices, and thats why its considered "good" not because it solves all problems. I believe that, the rest is ill-supported conjecture.
To point out the global stupidity that is economics, there is absolutely no consideration for the concept of a constraint. Constraints are a fundamental problem in optimization.
This article lays it out nicely. https://sloanreview.mit.edu/article/rethink-the-growth-imper...
Research this topic and you'll find economists arguing insistently that the conservation of mass or energy don't exist.
Welcome to economics! Bullshit and conjecture all the way down.
https://arxiv.org/abs/2405.17739
I'm finding this to be true with my experience. Noivces used to be a self solving problem... they got better or gave up.
Now they can present 80% solutions which cause bugs and eat up more experienced devs time. The worst part is they dont build the skills as they program continuing to defer to GenAI
Given constant weights / biases of a Transformer / DNN you could use pipelining to feed forward calculations through the array one layer at a time. For DNN's with thousands of layers you might see 1:1 speed up per layer channel.
I doubt they would undergo this process for marginal gains.
How do you imagine companies with staying power will be shaped in the future? Will we see new paradigms in management? ie smaller teams, jack of all trade types of individuals vs specialists, potentially the elimination of middle management all together
Ie founders create order out of chaos, then entropy slowly creeps in due to X,Y,Z?
I loved the lean startup I'll be reading this book as well :)
> Testing their solar-thermal desalination technique using samples of water from the Pacific, Atlantic, and Indian Oceans, Guo and his team were able to make the surface self-cleaning. In other words, it extracted freshwater and directed the remaining salts to the passive region where they could be later collected without reducing the panel’s efficiency.
This is not "large" this is a moderate improvement. Albedo is likely only marginally affected, and the solar power input over area is the same.
Depending on this cost of this process it could very likely be a wash in terms of NPV
LLM's require offline training and dont actually learn from their "lived / sessions / chats ect". Those can be used for training data but its not like its an implicit part of the technology.
For this reason I would say LLM's are not conscious, or automatically conscious.