adinhitlore

over the last months I've experimented with many alternatives to transformers, such as one i created github repo about:https://github.com/bggb7781-collab/lrnnsmddsArchitectures I've experimented with and my personal notes as pros and cons:1. RNNS: Like Mamba, RWKV and my attempt above: likely the best alternative to transformers but hard to parallelize and I've personally encountered weird...

So, not trying to "flex" but in the past I've worked on projects even for Microsoft/darpa, won awards, now been developing alternatives to transformers for generative models in C and can't even get an interview after tens of applications? even $10/hour seems like a stretch...looks like it's all about connections now and any buzzword or "complex code" is immediately considered an ai-slop or far...

OK, this is just social science but note this:Musk posted today for people to give 3 proofs of exceptional ability if they want to join the most expensive and "most innovative" company ever - Spacex.But how can you innovate when you follow the rules? what if you send 1, 2 or 4 reasons? and why would anyone work for someone with very questionable and controversial views? It doesn't seem possible...

Assuming it exists? in addition: if it's without phone app just windows app or is this super retro these days and everyone wants iOS and android? 10x.

So like chatgpt.com but on your mobile/computer without internet access necessary, $0, absolutely free? This isn't a question if it's possible but if it's a good idea? I will. 100% privacy and 0$, why not?

I was just watching coverage of the recent attack against Altman:https://www.youtube.com/watch?v=UJi4LJWC1zsmolotov cocktail against heavy guarded home seems almost harmless but what if they attack larger infrastructure, nuclear, terrorism? The attack vectors seem endless and very real...from hacking to actually using AI (deepfake) to make anti-ai content.

To be precise in 1903 they predicted it will take one million years in the best possible case (sarcasm i guess) for humans to build useful heavier-than-air craft. Well it took 70 days. The very same year the Wright brothers arrived or rather "left"...upwards with their...

or is just me? looks like jobs or personal real estate sales are failing me hard...if i was younger (now late 30s) i would've considered a fake firefighter stripper path, for now i'd persist on this cursed route.

https://raw.githubusercontent.com/bggb7781-collab/lrnnsmdds/refs/heads/main/lrnnsmdds

I just noticed it takes literally ~5 minutes to train millions parameters on slow CPU...but before you call Yudkowsky that "it's over", an important note: the main bottleneck is the corpus size, params are just 'cleverness' but given limited info it's powerless.Anyway, here is the project:https://github.com/bggb7781-collab/lrnnsmdds/tree/maincouple of notes:1. single C file, no dependencies....

So over the past months I've experimented with many potential alternatives to transformers, one of which is the so called Spiking Neural Networks which, to the ones unfamiliar with them, is an approach to AI which seeks to make AI as similar to the human brain as possible, in other words while transformers or RNN, LSTMs are result-driven (not seeking to simulate human 1:1) - SNNs actually try...

this is a linear RNN I've worked on, with the goal to solve the long range memory problem in RNNs as well as other improvements such as constant speed in generation...the rest of the text is copy/pasted as the original since due to the limited size this is massive overfit (perplexity is under 1.20):https://i.imgur.com/p6AmBrq.pngAnother way explain this is: "like Mamba but with good long term...

Hypothetically speaking a .bin model file <4 GB can be very competitive, the rest is just interface like a desktop app. This may be:1. free; 2. solves the privacy issue so you communicate with your chatbot offline, lack of privacy is huge issue amongst end users. 3. crazy but even in 2026 people don't always have internet everywhere.This turns the LLM industry from SaaS subscription to a...