@txbraindump

asking better questions

Joined December 2021
using large LLM to create synthetic data to finetune the small-LLM, gemma-4-e2b.
1
5
created a synthetic dataset for deferential calculus.
2
ankit retweeted
I bet we figure out that neural networks can be decomposed into evolved symbolic systems of sorts and that it will be achieved before capital S Superintelligence but everyone has to lock in
137
35
11
1,100
44,750
train for parallel search
12
ankit retweeted
Quantum sex, I mean entanglement.
22
235
17
1,577
27,160
tempering semetic neurons
11
conviction is gold
10
a bad things about having great memory is that u learn more and fast not only applies to good information but also for bad information
13
prediction intelligence - research control intelligence- engineering
3
23
need more RLHF for control guy
trying to balance
1
are u research guy or engineer guy
1
8
research ofc
2
better arch smart or better memory smart
23
self obsessed has least control
11
node which has biggest impact in chained information can consider himself as alpha
18
order from randomness creating new order
1
16
infinite can never have 1
1
4
we are in simulation
3