Errant
@errantv.bsky.social
noncanonical amino acids, in vitro display, and multispecific engineering
Thanks! Knew you'd be on top of how it affects PA
December 29, 2025 at 10:24 PM
Thanks! Knew you'd be on top of how it affects PA
December 29, 2025 at 10:19 PM
Ezra Klein's opinions are not worth $40k a pop. In fact Ezra Klein should have to pay ME to get to tell me his garbage opinions.
December 29, 2025 at 2:42 PM
Ezra Klein's opinions are not worth $40k a pop. In fact Ezra Klein should have to pay ME to get to tell me his garbage opinions.
TBH this is egregious enough that University of Richmond's ORI should investigate Lash.
December 29, 2025 at 2:11 PM
TBH this is egregious enough that University of Richmond's ORI should investigate Lash.
8-10 days till you're feeling normalish
December 29, 2025 at 12:48 PM
8-10 days till you're feeling normalish
WHY would you use a generative language tool that doesn't process information to search for information??? You're trying to find information not guess the statistically most probable sentence in response to a prompt.
December 28, 2025 at 10:49 PM
WHY would you use a generative language tool that doesn't process information to search for information??? You're trying to find information not guess the statistically most probable sentence in response to a prompt.
This is what search tools are literally designed for!! and they do the job better without hallucinations!!!!
December 28, 2025 at 10:48 PM
This is what search tools are literally designed for!! and they do the job better without hallucinations!!!!
Striking but not surprising, given that they're the population most ill-equipped to evaluate complex technologies.
December 28, 2025 at 9:00 PM
Striking but not surprising, given that they're the population most ill-equipped to evaluate complex technologies.
arxiv.org/abs/2409.05746
If you don't understand that hallucinations are mathematically impossible to eliminate or even meaningfully mitigate, then you don't really understand how LLMs work.
If you don't understand that hallucinations are mathematically impossible to eliminate or even meaningfully mitigate, then you don't really understand how LLMs work.
LLMs Will Always Hallucinate, and We Need to Live With This
As Large Language Models become more ubiquitous across domains, it becomes important to examine their inherent limitations critically. This work argues that hallucinations in language models are not j...
arxiv.org
December 28, 2025 at 8:55 PM
arxiv.org/abs/2409.05746
If you don't understand that hallucinations are mathematically impossible to eliminate or even meaningfully mitigate, then you don't really understand how LLMs work.
If you don't understand that hallucinations are mathematically impossible to eliminate or even meaningfully mitigate, then you don't really understand how LLMs work.
LLMs aren't search! They are statistical prediction of language! All they do is predict the most likely words to come next in response to a prompt! There is no information processing happening!
December 28, 2025 at 8:44 PM
LLMs aren't search! They are statistical prediction of language! All they do is predict the most likely words to come next in response to a prompt! There is no information processing happening!
False results != fabricated results as you are GUARANTEED to get when using LLMs
December 28, 2025 at 8:43 PM
False results != fabricated results as you are GUARANTEED to get when using LLMs
Using a language tool that will WITH MATHEMATICAL CERTAINTY generate false information at a predictable rate is NOT perfectly fine for legal research JFC
December 28, 2025 at 8:37 PM
Using a language tool that will WITH MATHEMATICAL CERTAINTY generate false information at a predictable rate is NOT perfectly fine for legal research JFC
Did any of these guys show back up today?
December 27, 2025 at 2:32 AM
Did any of these guys show back up today?