> What service does ham radio actually provide to society?
Ham radio allows the public to use slices of the radio spectrum.
If someone wanted to "hack" on RF projects, what should this person do if the entire spectrum was given to companies and governments + militaries instead?
If you are arguing against ham radio you are basically arguing against hacking, tinkering and experimenting.
It's true, this will certainly not be a high speed network as 2.7 kHz is the max allowed bandwidth (compared 500 kHz for LoRa). But it should be fast enough for transmitting text messages.
Antenna size is problematic, but NVIS does not require the antenna to be high up in the air. Also the polarisation must not be vertical, so throwing a simple wire dipole on the ground, might actually do the job. But we will see, it's an experiment.
> The mostly unpredictable and extremely low duty cycle modes of non-line of sight VHF/UHF propagation like tropospheric ducting are not a realistic option for communications networks. Alternately the more reliable tropospheric scatter is a brute force solution requiring high output power of a kilowatt or more with big horn antennas.
A friend and myself are currently prototyping a solution where we transmit data using the near vertical incidence skywave (NVIS). Possibly the only option, when you want to avoid infrastructure at all costs. Of course you have the disadvantage of huge HF-antennas. To make this setup usable at all, we are trying with 20m long copper cables close to the ground. If this does not work we will try magnetic loop antennas.
> Humans "hallucinate" in the AI sense (it's an awful word that obscures how often we do it) all the time too.
Agreed. I'd like to add another point to the discussion. It seems to me, as if LLMs are held to a higher standard regarding telling the truth than humans are. In my opinion the reason for this is that computers have been traditionally used to solve deterministic tasks and that people are not used to them making wrong claims.
> I don't understand the "coincidentally" argument.
Nothing is coincidental about those models. They were designed after processes in the brain. They underwent rigorous training to generate a function that probabilistically maps inputs to outputs. Eventually, it exceeded the threshold where most humans consider it to be intelligent. As these models grow larger, they will surpass human intelligence by far. Currently, large language models (LLMs) have fewer weights than human brains, with a difference of a factor in the thousands (based on my superficial research). But what happens when they have an equal or even 100,000 times more weights? These models will be able to model reality in ways humans cannot. Complex concepts like the connection between time and space, which are difficult for humans to grasp, will be easily understood by such models.
> LLMs do not hallucinate sometimes. They hallucinate all the time, it just is a coincident that sometimes these autocompletion of Tokens aligns with the reality. Just by chance, not by craft.
That is such a weird way to think about them. I'd rather say, they always provide the answer that is most probabilistic according to their internal model. Hallucination simply means, that the internal model is not good enough yet and needs to be improved, which it will.
The fundamental concept behind LLMs is to allow the model to autonomously deduce concepts, rather than explicitly engineering solutions into the system.
This paper highlights a crucial aspect of evaluating AI language models: the significance of prompt construction (e.g. adding "think step by step").
When a model is given insufficient context beyond the question, it may generate responses based on its best guess. This situation can be compared to abruptly waking someone up in the middle of the night and demanding an immediate response to a question.
In contrast, when humans are asked to answer questions in a test setting, they are aware of the larger context and the importance of providing accurate answers.
> Because of my own personal philosophy regarding technology and AI, all the code in this repository that was written by me - I wrote 100% on my own. There is and will be no usage of Github Co-Pilot or any other AI tool. I became a software developer because of my passion for our craft - Software Engineering. I build this tool because I enjoy programming. Every single line of code you'll read in this repo, that was written by me, is produced first in my mind and then manifested into reality through my hands. I encourage any contributor to follow the same principle, though I can't and don't want to put any restrictions on this. Just like people stopped walking because they commute by cars and trains, which caused an increase in obesity and illness, I believe that the massive usage of AI will cause people to stop thinking and using their minds and the resulting havoc is unthinkable.
Not everyone is trolling on the internet my fellow human.
> This is why I am pretty sure you're trolling. This is a lot like saying "You can make GPT-4 do prime factorisation of a semiprime by noting the constituent primes." The point of testing the model is to see if the model can give you the answer, not to give it the answer.
It is not "giving the answer to the model" if you adjust the prompt with additional instructions like "this is a test question" or "use common sense". It is about creating the correct frame of context. With these kind of adjustments GPT-4 passes every novel Winograd scheme I have presented it with.
> He broke in to an MIT networking closet (he was never a student there) and connected his equipment to the network.
The closet was unlocked and he used a regular guest access to the MIT network. Also he was downloading documents that were created by using public funds.
> There are a lot of much more legal ways to make the Internet freer. He was a smart guy and knew what he was doing was highly illegal.
There are always other and more effective ways to everything. With this kind of argumentation one always must come to the conclusion that it is best to do nothing. Also let's not forget that he did much more than downloading documents at MIT.
> think that this is the major issue with martyrdom. Aaron is remembered for "fighting the man" but the real story is a significantly muddier than that. A martyr's death makes it seem like the martyr did nothing wrong even if they did.
That's a definition for martyrdom I have not heard before. Usually a martyr is simply defined as a person who is willing to suffer or even die for a cause, belief, or principle that they consider to be of great importance.
> Sorry, I know this is kind of a dumb and not so productive soap box. Oh well.
I will simply never understand why people will argument so strongly against their self interests.
> But it's incorrect to say that they pass every test.
Interesting! So there is published evidence, that it cannot pass novel Winograd schemas? Could you please provide me with the sources? I would like to have a look at the prompts. From my experience you can make GPT-4 pass novel Winograd Tests by giving GPT-4 additional context.
> I think not much thought was given to explain the answers in depth since the main point was to show where GPTs reasoning was lacking.
Funnily enough it seems as if the prof has also fallen for a reasoning error then.
How can both be equally true? The line gets longer than 100 people and the social cost of the coffee is exactly 100 dollar (1 Dollar for every person).
> Question 2: In a country where everyone is identical, 100 people wait in line each day to buy raspberries at a controlled price. The government has decided to hand out free coffee to the people standing in line. The coffee costs the government $1 per cup, but the people in line value that coffee at only 75 cents per cup. What is the social cost of providing the coffee?
> Grading Remarks: This answer completely misses the key fact that free coffee will cause the line to get longer (in fact it must cause the line to get longer, given the stated assumption that everyone is identical, hence initially indifferent between standing in line and not standing in line). In fact (unless one assumes a very small population), the line must grow until the extra waiting time completely dissipates the value of the free coffee; thus the social cost of providing the coffee is $100.
Wait what? So the line must get longer, *given the stated assumption that everyone is identical*. Well what about the given stated assumption *that 100 people wait in line each day*? One of the stated assumptions can be dropped just like that, while the other is treated like a law of nature? Also where is the logic in "everyone is identical, hence initially indifferent between standing in line and not standing in line"? Am I missing something?
I am pretty sure that the development of a consciousness in coming LLMs is unavoidable. A consciousness is just another useful abstraction for making precise predictions about the world.
The big question will be: How do we treat machines that have a consciousness? Is consciousness in itself worth protecting, or is it the human attributes (being able to feel pain and an evolutionary priming towards survival) that should be granted this special status. This is going to be a fun discussion.
> the real surprise is that language conveys concepts better than we previously thought?
Is it really that much of a surprise? Isn't the whole purpose of language to transport concepts?
I mean, our brains are not directly connected to each other, yet you just transferred a concept (which was a result of your thinking and understanding) to my brain by using language.
I've been thinking a lot about the ability of neural networks to develop understanding and wanted to share my perspective on this. For me it seems absolutely necessary for a NN to develop an understanding of its training data.
Take Convolutional Neural Networks (CNNs) used in computer vision, for example. One can observe how the level of abstraction increases in each layer. It starts with detecting brightness transitions, followed by edges, then general shapes, and eventually specific objects like cars or houses. Through training, the network learns the concept of a car and understands what a car is.
The same principle applies to Transformer networks in text processing. Instead of pixels, they process textual elements. Neurons in different layers learn to recognize complex relationships and understand abstract concepts.
Ham radio allows the public to use slices of the radio spectrum.
If someone wanted to "hack" on RF projects, what should this person do if the entire spectrum was given to companies and governments + militaries instead?
If you are arguing against ham radio you are basically arguing against hacking, tinkering and experimenting.