Nina
@NinaPanickssery
Currently full-time mothering. Interested in AI safety & interpretability.
I think writing a PhD thesis on smell is a very good idea and I would be tempted to do the same It would look something like this:
underrated effect of LLMs: releasing new open source library updates/versions will be much more of a nightmare because all the models will keep suggesting your old API and get confused by newer code
Owain+collaborators are a machine!
New paper & surprising result. LLMs transmit traits to other models via hidden signals in data. Datasets consisting only of 3-digit numbers can transmit a love for owls, or evil tendencies. 🧵
Data augmentations from Benjamin Franklin gutenberg.org/files/20203/20…
Just got invited to take part in an expert elicitation study about generative AI's impact on cognition and education and the invitation letter had THIS GODDAMN LINE.
on grok.com, the backend sends the full (not summarized) CoT to your browser. it's not displayed in the UI, but you can see it with browser dev tools or w/e check out the json payload of responses from `grok.com/rest/app-chat/…{conversation_id}/load-responses`
chain-of-thought monitorability is a wonderful thing ;) gist.githubusercontent.com/nostalgebraist…
some people must be in the market for this right?
do any rationalists with short timelines want to borrow money from me at usurious rates?
The received wisdom is that LLMs are more useful for frontend dev compared to backend dev. But having used them for both, I'm suspicious. Curious what people who do full stack dev for prod applications with real users think. x.com/esrtweet/statu…
I think I qualify as an experienced open source developer, and that study looks completely ridiculous to me. I've discussed it with some peers. We think one of the confounders may be that LLMs are much better at accelerating green-field development then fixing or improving large…
Futurist forecasting is funny because it's hard enough to determine what exactly is going on right now, at this very moment, or in e.g. the past five years (which are notably in the past and already happened), let alone what the world will look like in 20 years or even later