monkeycantype

no profile record

comments

5 месяцев назадdiscuss

One of my favourite conversations I ever had with another human being, was with someone who was telling me about a book he was reading, which he evidently loved very much. He spoke almost no English, and I spoke absolutely no Cantonese, so the entire fifteen minute conversation was conducted with gesture, intonation and the phrases: 'one man' 'many mans' 'some kung fu' 'many _(* n)_ kungfus!'

I love the left hand of darkness, the dispossessed, all the hainish books, but I have never loved any book as much as my conversation partner loved his book about many many Kung Fus.

5 месяцев назадdiscuss

I only know your mother from a distance, although sometimes that was only few blocks apart on the map, I'm not aware of even passing her on the on the street, I knew only the part of her revealed in her writing and records of her public speaking, but even from that distance she has been one of the most influencial people in my life, her writing had deep impact on me, my view of the world and the person I have become.

8 месяцев назадdiscuss

yes I agree it's not the angle of the article, but it is my entry point into the idea/concern/unanswered question at the end of the article “My worry is not that these models are similar to us. It’s that we are similar to these models.” - that the enormous difference in the medium and mechanics or our minds and llm's might not be that important.

before i go any further, let me first reference The Dude:

     - "this is just like, my opinion man."

I’m down with the idea that LLM’s have been especially successful because they ‘piggyback on language’ – our tool and protocol for structuring, compressing, and serialising thought, which means it has been possible to train LLM’s on compressed patterns of actual thought and have them make new language that sure looks like thought, without any direct experience of the concepts being manipulated, and if they do it well enough we will do the decompression, fleshing out the text with our experiential context. But I suspect that there are parts of my mind that also deal with concepts in an abstract way, far from any experiential context of the concept, just like the deeper layers of a neural network. I’m open to the idea, that just as the sparse matrix of an LLM is encoding connection between concepts without explicitly encoding edges, I think there will be multiple ways that we can look as the structure of an AI model and at our anatomy so that they are a squint and a transformation function away interesting overlaps. that will lead to and a kind of 'god of the gaps' scenario in which we conceptually carve out pieces of our minds as, 'oh the visual cortext is just an X', and deep questions about what we are.

8 месяцев назадdiscuss

Until we have a testable, falsifiable thesis of how consciousness forms in meat, it is rash to exclude that consciousness could arise from linear algebra. Our study of the brain has revealed an enormous amount about how our anatomy processes information, but nothing of substance on the relationship between matter and consciousness. The software and data of an operating LLM is not purely abstract, it has a physical embodiment as circuits and electrons. Until we understand how matter is connected to consciousness, we also cannot know whether the arrangements and movements of electrons meet the criteria for forming consciousness.

9 месяцев назадdiscuss

Hello Rohan. This is really great. If you are able to include parameters to expose the intermediate data as inputs and outputs, so that this can be run to a step in the process and output the data, or run from a step with pre-prepared data. It would mean that other people could build on what you've done to create other diagrams and renderings.

10 месяцев назадdiscuss

as I dabble with neural networks, I keep having these moments where i wonder, is this what I am? (a neural net) And it has begun to make we wonder in an entirely unscientific manner, whether a large part of what notice as neurodivergence, is not the core divergence from 'typical', but something emergent from that difference, that we are noticing is that we are interacting with a mind that has trained with an uncommon loss function, on different features of reality. There is only some much space in one head and so depending on what apsects of reality we are drawn to poetry,football,horseriding,music,art,software,cooking,farming,other people we end up very different people

10 месяцев назадdiscuss

no, chi and centrifugal force do not really exist. But because everything we percieve is essentially a metaphor, a model of reality, sometimes chi or loosy goosey uses of the word 'energy' really can be a valid heuristic for the things that you can do to optimise your 'ability to do work'. As long as you remember its a vague, subjective, context restricted heuristic, and don't try shooting chi bullets, I say max your chi flow friend.

10 месяцев назадdiscuss

I think it's the theory in the article linked just above by renewiltord

10 месяцев назадdiscuss

Ah! That's the thread, not the source I read but I beat if I read the referenced paper I get there , thank you

10 месяцев назадdiscuss

I wasn't but that's intriguing Thank you

10 месяцев назадdiscuss

no there is a very specific reason, related to mapping the 2d surface of your body to a 2d mapping on your brain that allows the areas of your brain that process sensory input from your skin to be adjacent to the processing of the areas that are adjacent on the skin that only works with a flip, I can remember what that is, I only remember the tingle of understanding it at the time

10 месяцев назадdiscuss

can you be my collective memory for a minute, I remember the existence of a very satisfying engineering explanation for why the representation of various body parts needs to be flipped left/right in the brain that came down to the topology of the wiring, and explained why unflipped isn't feasible / or perhaps it was just less efficient, it was one of those 'mind explodes' moments, but now I can't recall the logic.

4 года назадdiscuss

I enjoy pretending that I would.

4 года назадdiscuss

Answering personally, I seem to be 'the swing voter' demographic for this phone - I have gone to their site many times, but have never paid them a dollar. I'm interested less because of what I could do with this phone - apple works great for me - but because I'm interested in how much the hardware can be open sourced, can a phone or laptop be built where only the low level components are sourced from existing manufacturers, without needing google or apple etc to curate the assembly. I like the idea of having what I have with apple, but not having apple or google or anyone own the ecosystem.

5 лет назадdiscuss

You’re welcome

5 лет назадdiscuss

I use dot, and manually position everything, then use the neato -n2 option to render:

`dot -T ${format} -n${n} ${verbose>1?"-v ":""}-Goverlap-true -Gsplines=false -Kneato -o "${ output }" "${ input }"`

this still doesn't do subgraphs well,

so I sometime do the 'groups' in a separate document, then composite the layers with image magick

`magick composite ${verbose>1?"-verbose ":""}-gravity NorthWest "${a}" "${b}" "${png_output}"`;

5 лет назадdiscuss

but it's solving one part of the problem, and doing a good job at it. I've been making diagrams using graphviz, but setting the positions and edge curves manually, pulling the nodes and edges out of a list in excel

https://github.com/mathew-j-davis/boxesandarrows

I've been setting the 'waypoints' for more complex curves (curves with many bends) manually (typing the bezier numbers in by hand), while I try and figure out how I want them routed, you're right it's not easy figuring out rules for where edge should go, let alone implementing it