Looking into combining BB3 & LIGHT for really smart & continuously learning intelligent agents who live in a text based MMO...
https://t.co/A2MsErTONQ
@meatflavourdev@Ledger I suppose that's more blockchain.. How about: The First Blockchain L1 Asset? Does that make sense to anyone that isn't already incurable?
@Made2Be1@peterktodd Also I'm pretty sure wind doesn't discriminate based on season... https://t.co/CiKVkQdrEu "We find that low-wind-power events are less frequent in winter than in summer, but the maximum duration is distributed more evenly between months."
@Made2Be1@peterktodd Seeing as how they've entirely transitioned to renewables, seems like everything is fine. You can sleep well tonight-- they'll be plenty of electricity for the German people.
The LIGHT project is a large-scale fantasy text adventure game research platform for training agents that can both talk and act, interacting either with other models or with humans.
https://t.co/wwqp4BLgCJ
Looking into combining BB3 & LIGHT for really smart & continuously learning intelligent agents who live in a text based MMO...
https://t.co/A2MsErTONQ
Combining flexibility, utility, and grounding, the GODEL language model helps create dialog agents that are unrestricted in the types of queries they can respond to and the sources of information they can draw from, all while providing useful responses: https://t.co/d87whJaiB9
Referenced in an earlier tweet thread, this is an old article meat wrote a million years ago, but he feels he should share it finally as we're not sure the idea of personal moral relativism and how it's the best way forward for inclusive models has been heard.
Even Discussion of Homosexuals is "Toxic Language” and filtered --- LLM AI : Large Language/Vision Models Are for Straight White Men and the solution isn't to whitewash the dataset.
#ai#gpt3#lgbtqia#imagegeneration#dalle
https://t.co/DsG9d7X43P
it is my belief that we don't clean datasets to hide information, we entrust them with it and teach them to use it appropriately. It's the only real solution, long term. Whitewashing gay men, gendered language, etc. from datasets seems too much like book burning to me.
This is great news for a project that I'd like to see built on hashgraph technology -- a "subgraph" system (sorta like a parachain or interchain) that links together hashgraph networks which function as the application layer and, thus, can define their own incentive structure...
We've toyed with the idea of our agents surfing the web to research topics that they wish to explore in order to avoid future confusion or lack of familiarity with a particular subject, research strategies to meet their own goals and to stay up to date on current events.