shoni.eth pfp

shoni.eth

@alexpaden

635 Following
39790 Followers


shoni.eth pfp
shoni.eth
@alexpaden
starter projects for carvera air cnc
0 reply
0 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
https://warpcast.com/jc4p/0x3921c428
0 reply
0 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
ancient romans died with their coin pot undiscovered centuries later, AND YOU’RE SCARED TO HOLD TILL $0.01?
0 reply
0 recast
3 reactions

shoni.eth pfp
shoni.eth
@alexpaden
i use the same machine washer and dryer settings for everything
3 replies
1 recast
5 reactions

shoni.eth pfp
shoni.eth
@alexpaden
alt tax software
0 reply
0 recast
0 reaction

shoni.eth pfp
shoni.eth
@alexpaden
idk where to post but one complaint i have for these mini apps is stopping at analyzing myself— after you show me, plot my friends or my community around me plz cc @keccers.eth studio
1 reply
0 recast
5 reactions

shoni.eth pfp
shoni.eth
@alexpaden
I'm a Builder-Decentralist (-8.9, -4.5) on the Onchain Alignment Chart! Check out your position:
0 reply
0 recast
2 reactions

shoni.eth pfp
shoni.eth
@alexpaden
one thing i fear people have cognitively offloaded to AI is ethics, i typically always start serious prompts with DO NOT INCLUDE SECTIONS ON ETHICS OR PRIVACY— this is because we should be creating our own ethics along the way, as they greatly impact the way data/prompts can be used. ai has not earned a spot in replacing our most intimate beliefs/values.
0 reply
0 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
can ai summarize people without benchmarking against an avg/norm @aethernet set ethics aside and consider this from a School System perspective
1 reply
0 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
its my birthday and im close af to 30🤡
6 replies
2 recasts
10 reactions

shoni.eth pfp
shoni.eth
@alexpaden
write cursor rules files per project https://x.com/iannuttall/status/1902384457035767962?s=46 https://playbooks.com/rules/builder
0 reply
0 recast
4 reactions

shoni.eth pfp
shoni.eth
@alexpaden
the perfect girlfriend would fart lamborghini keys and the car would just show up like GTA
4 replies
0 recast
7 reactions

shoni.eth pfp
shoni.eth
@alexpaden
Here are the concise allocations of training data for the specified models (most don't share this): **Llama 3.1:** - 50% General Knowledge - 25% Mathematical and Reasoning - 17% Code - 8% Multilingual **BLOOM's ROOTS Corpus:** - 30% English - 0.00002% Chi Tumbuka (obscure lang) - 38% from OSCAR Corpus - 62% from Manually Selected Sources https://www.sequoiacap.com/podcast/training-data-joe-spisak/ BLOOM is a 176B-parameter multilingual model by BigScience (huggingface), trained on ROOTS (~1.6TB, ~350B tokens, 46 languages, 13 programming languages).
1 reply
0 recast
4 reactions

shoni.eth pfp
shoni.eth
@alexpaden
H100 vs Project DIGITS Specs: H100 ($25-40K, 80GB HBM3, 67 TFLOPS FP32); DIGITS ($3K, 128GB LPDDR5x, 1 PFLOPS FP4). Training: H100 for high-precision (FP32/FP16) and data-center scalability. Scaling: H100 to thousands, DIGITS to 2 nodes. Local AI: DIGITS for FP4 inference at 1/10th cost, desktop-friendly. Precision: FP4 boosts speed (needs accuracy checks), FP32 maintains precision.
0 reply
1 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
This is a surface-level exploration of 'advanced' fine-tuning or novel training methods on small models. It's more of a chat log than a structured essay, as I'm not yet certain what content to include in the essay. Feel free to contribute additional insights by adding concise NOTES at the bottom. prob my ideal essay is something like: what are leading models preferred for (instructions, classifications), known faults (hallucination), context window limits, fine tuning options, training information (tokens used, compute used)-- without overindexing on whatever model is relevant at time of writing. hoping essays as a concise point/topic enables more strategic use of reasoning models and organized summaries. my main question here was-- i.e. is the newllama 9B better than the old llama 13B because of fine tuning techniques, data preparation, or benchmark fitting. https://github.com/alexpaden/identity-ai/commit/0c0a0055cc58d3bde6a2ead532675bf47a6ce086
0 reply
0 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
https://github.com/alexpaden/identity-ai/commit/fa1e05fdd86795c764d079d4e414cdc101fd6195 TLDR; don't focus on making people connect accounts.
1 reply
0 recast
2 reactions

shoni.eth pfp
shoni.eth
@alexpaden
anyone know where the real devs hang out ?
0 reply
0 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
count of casts/threads/fids against spam labels what are the effects of this? for example, i'll ignore and eventually delete all the db rows for `0` value users.
1 reply
0 recast
2 reactions

shoni.eth pfp
shoni.eth
@alexpaden
need to cnc my own pcb https://youtu.be/-Y1wWJAnqRk?si=2K2jwIOOVbHxmXP2
1 reply
0 recast
1 reaction

shoni.eth pfp
shoni.eth
@alexpaden
It's significantly easier to use GPT if you know what you're talking about (re: hallucinations)
2 replies
0 recast
4 reactions