Iโm excited to try Gemini Neon.
Its capabilities in cutting-edge frontier work should become apparent very quickly.
Modelsโ tendency to regress towards the mean and their lack of comprehensive understanding of novel systems have been major challenges in development. These are the limitations Neuromorph Nano is designed to address.
Currently, development requires extensive manual tuning because Anthropicโs and OpenAIโs models have plateaued in their ability to advance the architecture without compromising existing components.
Frontier modelsโ limited comprehension and tendency to regress towards the mean of their training data remain significant obstacles to novel development.
These consumer-facing frontier models are not particularly well suited to developing frontier technologies. Perhaps this reflects a fear of competition, or perhaps it is a limitation of their underlying designs. It is difficult to say.
We can advance independently.
35
Being broke as fuck is actually awesome. It gives you a valuable perspective and encourages you to be creative with limited resources.
Iโm genuinely going to cherish this time in my life. You learn a great deal about the people who stand by you when you have nothing to offer them, at least not at the moment, and perhaps never.
However, my time is running out. I have another week or two before I really need to produce something that can pull me out of the hole Iโm in.
25
Neuromorph NANO should be finished by the end of the month, ideally.
Iโm not sure whether to make it source-available under a non-commercial licence and license out commercial use separately, or keep it closed-source and proprietary.
It wonโt be easy to reverse-engineer without access to the source code, at least given the current state of models. But who knows what will be possible in the future.
I do want exceptionally talented, intelligent, and creative collaborators to help develop it further.
The system has potential and may influence the direction of AI engineering and architecture in interesting ways, though only time will tell.
It would be great to allow some genuinely brilliant minds to contribute a few PRs.
21
GAVRIEL.TECH retweeted
For the past few days, I've been trying to figure out the right terrain texture for M.E.C.'s final renders and animations, and it's proving to be quite a challenge!
I'm torn between procedural textures, which could be extremely heavy to render, and image-based textures, which might look repetitive or lack detail in close-up shots.
This is the terrain I currently have. I'll be doing more test renders while experimenting with different textures, lighting, and materials to find the right look.
One thing I'm certain about is that I want the planet's surface to remain dark, rather than the typical brown or sandy terrain.
If anyone has experience with creating realistic planetary surfaces in Blender, I'd really appreciate any tips or suggestions!
Still figuring this one out! :)
GAVRIEL.TECH retweeted
I got distracted a bit by adding a full circuit board simulator, 3d renderer, including custom firmware compiler for esp32 s3,c3 and AVR and allowing it to have a 3D printed case with lightpipes. Fab is going to do it all
There is still a great deal of work left to be done in ML. If you cannot see that, you have a limited imagination.
46
I havenโt left my apartment in weeks. All I do is work on this ML project, from the moment I wake up until I go to sleep.
This is the same pattern of behaviour that caused my wife to leave previously, and I cannot really blame her.
It is also very cold outside.
At this point, I feel compelled to conserve whatever energy I have for this work and whatever money I can spare for food to feed us. We ran out of almost everything last night.
All we have left are three eggs, one piece of toast, about 50 grams of pyttipannu (which my wife hates), one tortilla, and some old porridge from the cupboard.
I can go several days without eating when I am deeply absorbed in work. I can get by on cans of tuna for days.
Of course, it wasnโt always this way. These are the leanest times we have experienced in many moons.
I can tell my wife is reaching her limits.
Lately, I have been working for twenty hours straight or longer.
It really isnโt that difficult for me.
The risk is that, even after completing the project, there may still be no profit or employment to show for it.
It makes me wonder whether the work itself, and the challenges inherent in these circumstances, are the rewards I am ultimately seeking.
Iโve become a gambling man just like my father, grandfather, and great-grandfather. Weโve just gambled in our own ways. I was never interested in Las Vegas casinos, as they seemed to be. My interest was always in technology, and technology alone. Even from within music.
56

























