vanishingideal pfp

vanishingideal

@vanishingideal

1099 Following
125 Followers


vanishingideal pfp
vanishingideal
@vanishingideal
gn
0 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
https://x.com/lukOlejnik/status/1898836732918788553
0 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
I think 3070 might be down bad.
0 reply
0 recast
4 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
Chunking by FID might have the benefit of isolating failure. Hope it doesn't cost an arm on lambdalabs. Cleaning data and removing stop-words might lead to better semantic results.
0 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
You were probably the earliest bid with size. Do you have any thoughts on this? https://warpcast.com/vanishingideal/0x3f44a305
1 reply
0 recast
2 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
How are you cleaning casts? How are you approaching chunks? Only seeing a hash/embedding pair here.
1 reply
0 recast
1 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
Here's a notebook to help you get started. https://github.com/mixedbread-ai/binary-embeddings/blob/main/mxbai_binary_quantization.ipynb
1 reply
0 recast
1 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
Have you considered int8 embeddings? https://huggingface.co/blog/embedding-quantization
1 reply
0 recast
4 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
I think DRB is this cycle's WIF, but fail to see what this ripples out into. Is this the run up to self-regulating AI with near-human agency? What does a step change from here look like?
0 reply
1 recast
2 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
I wonder where the line is, from choosing to be mildly inefficient to becoming lobotomised.
1 reply
0 recast
2 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
A warm welcome @frogprog. He's probably my favourite educator/content creator in the area of smart contracts. You can check out some of his work here: https://www.youtube.com/@smartcontractprogrammer vyper-by-example.org solidity-by-example.org www.smartcontract.engineer Any advice for him @horsefacts.eth @hellno.eth @rish @cassie?
0 reply
1 recast
5 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
Which text embeddings model are you using?
1 reply
0 recast
3 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
gm
1 reply
0 recast
3 reactions

vanishingideal pfp
vanishingideal
@vanishingideal
Prompt-engineering is certainly a skill, but there's tacit knowledge around using Google to find things. Don't be quick to give that up.
0 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
Sure, but don't they use search engines for real-time information?
1 reply
0 recast
1 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
Why?
1 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
Pretty simple technique for self-aggrandizement is to signal humility.
0 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
https://www.danlj.org/mkj/lad/contents.html
0 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
If you were ever wondering how mentally stable linux programmers were.
1 reply
0 recast
0 reaction

vanishingideal pfp
vanishingideal
@vanishingideal
Accepting both simplifies things. Incentives for $QR i.e., a 10% bid boost vs. ETH or some post-auction incentive for winning bids made in QR where they get some of it back might be interesting to A/B test.
0 reply
0 recast
1 reaction