If you're interested in my work on the intersection of LLM/ontologies/biology, head on over to linked in or mastodon where I post more, including some recent work:

DRAGON-AI manuscript now published on JBMS!

RT @GBBranstetter:

No, I don’t want to know how agents leverage knowledge to help customers. I want you to tell me when my delivery will be here!.

Hey @Sunbasketmeals your delivery company forgot to configure the RAG on their rubbish AI chatbot

Nice explainer on @news4go from @prmshra !.

RT @ontowonka: Benchmarking #variantprioritzation tools for #raredisease diagnostics is critical! See @MonarchInit new eval framework at: h….

RT @WiringTheBrain: 𝗧𝗵𝗲 𝗚𝗲𝗻𝗼𝗺𝗶𝗰 𝗖𝗼𝗱𝗲 - 𝘁𝗵𝗲 𝗴𝗲𝗻𝗼𝗺𝗲 𝗶𝗻𝘀𝘁𝗮𝗻𝘁𝗶𝗮𝘁𝗲𝘀 𝗮 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗺𝗼𝗱𝗲𝗹 𝗼𝗳 𝘁𝗵𝗲 𝗼𝗿𝗴𝗮𝗻𝗶𝘀𝗺 🧬.very excited to….

Hey @Docker, what's with the sudden revoking of our sponsored open source subscription? We are getting this for @linkml_data and colleagues getting the same thing for @OBOFoundry. 🙏 for the O/S subscription, but more advance warning of its cancelation would have been nice☹️

RT @MonarchInit: Twitter/X no longer aligns with our values; we are intentionally choosing to communicate about our work using other media.….

RT @Bioconductor: **We are moving platforms**. Bioconductor will cease posting on Twitter/X as of December 17, 2023. Instead we will be con….

RT @leechuck: New work by my former student (now Prof) @azzathamer on phenotype-based prioritization of genetic variants for #RareDisease T….

It looks like there are already a few products out there already targeted at this use case. Curious if anyone has tried them? Do they offer analogous functionality to sudowrite StoryEngine for fiction generation

I think the headline is over-egging a bit, there are definitely things that could be improved about the current system, but ultimately it’s necessary to communicate our vision through language and visuals, and use of AI will inevitably grow here, and we need to adapt to this.

The whole thing gets silly when reviewers start using chatgpt to summarize proposals written by chatgpt… (I would never use it for this, and only feed publicly available docs over APIs, but I’d be surprised if this isn’t happening).

I can see the appeal of an end to end solution, especially for boilerplate docs. Start by feeding in papers with preliminary work, orcids of team, converse about aims outlines, have it deal with retrieval, summarization, consistency, ….

I personally use chatgpt occasionally to generate initial drafts for initial discussion, but the generated texts rarely makes it into final version. Anecdotally, some have more success, some less. Some have developed elaborate prompts.

I think a variant of sudowrite for writing grant proposals would do well (not saying this is a good thing, but likely inevitable given effort to payline ratio). [article: ChatGPT use shows that the grant-application system is broken] 🧵.

RT @BioDataJournal: 🙌Closer to bridging the gap between #biodiversity #omics data standards!.🆕Methods paper - part of the @Bicikl_H2020 col….

RT @linkml_data: The next LinkML Community meeting is tomorrow, Thursday, Sept 14 (the second Thursday of the month, rather than the usual….