Joe Golden

Current: Listen to your documents.
Prev: Co-founder/CEO. UMich PhD econ. Google / Microsoft / Upwork.

[Joe Golden @JoeGoJoeGo]

@TheStalwart Yes, rerun the whole thing. Flash 2.5 is generally better than Flash 2.0. So is Flash 3.1 Lite, which is less expensive. I've been using these models for my startup and would be happy to help you with this project. We follow each other and have the same name!

[Joe Golden @JoeGoJoeGo]

@ModeledBehavior Bad idea! I refused to learn cursive in 3rd grade and learned to type instead. I stand by that decision.

[Joe Golden @JoeGoJoeGo]

@LangburdWright @bocowgill @pablhern I really enjoyed reading this paper! The experiment made sense, the results were interesting, and the writing quality was very high.

[Joe Golden @JoeGoJoeGo]

@lukestein @johnjhorton @apo_filippas Legally mandated reputation inflation!

[Joe Golden @JoeGoJoeGo]

@chi_t_williams @KelseyTuoc It's about extracting various types of signals, including style, topics and potentially PII. They even run text through an LLM to try to disguise it, not entirely successfully. I'm now convinced that if you have enough public and anonymous online writing, LLMs can dox you.

[Joe Golden @JoeGoJoeGo]

@tab_delete I'm not sure what pickups are unless they're in a guitar, but it sounds like you recorded your own book, which I always appreciate!

[Joe Golden @JoeGoJoeGo]

@chi_t_williams @KelseyTuoc This paper I read https://t.co/4svs8fMAiT suggests identification isn’t about lots of your text, it’s about whether a stable “style signature” can be inferred. Even limited writing can be enough once those statistical patterns show up.

[Joe Golden @JoeGoJoeGo]

@aviel Caltrain wifi is working again? Source: it wasn’t one on of my 2 trips on it last week.

[Joe Golden @JoeGoJoeGo]

@johnjhorton Did your son post this note at the factory he owns?

[Joe Golden @JoeGoJoeGo]

@johnjhorton Is it safe to assume you’ve never been to space?

[Joe Golden @JoeGoJoeGo]

@johnjhorton I didn’t know that you moved to an apartment above a bowling alley and below another bowling alley.

[Joe Golden @JoeGoJoeGo]

@pepsi Rum and coke

[Joe Golden @JoeGoJoeGo]

@Noahpinion Amazingly, they used Delve for their SOC compliance hahaha

[Joe Golden @JoeGoJoeGo]

@jt_kerwin Haha they must have trained on Futurama scripts.

Lucy Liu: "I'll never forget you, Fry. MEMORY DELETED."

[Joe Golden @JoeGoJoeGo]

@aviel I absolutely love this idea and would love to be involved.

[Joe Golden @JoeGoJoeGo]

@FederalAsh14 @johnjhorton @MeganTStevenson Qianfan OCR looks interesting. Here's what's different about it from other OCR models. It is a bigger model, but it can also handle, document understanding and key information extraction. It takes a prompt, and uses thinking tokens. I listened to the paper with Paper2Audio!

[Joe Golden @JoeGoJoeGo]

@FederalAsh14 @johnjhorton @MeganTStevenson I'm starting to look into Qianfan OCR as well. It is 4B parameters, but claims to go beyond the other recent "OCR" models to handle document understanding tasks (like chart summaries). It looks interesting and I'm going to read its paper this week too.

[Joe Golden @JoeGoJoeGo]

@lukestein @boyuan_chen @nberpubs @NotebookLM @Paper2Audio Thanks for continuing to share Paper2Audio! Please share feature requests or other feedback anytime.

[⑆Luke Stein⑈ @lukestein]

@JoeGoJoeGo @ben_golub @BarbaraBiasi Late reply but I too learned about @Paper2Audio from @ben_golub too, and it's so good. Thank you, Joe!!

[Joe Golden @JoeGoJoeGo]

@MeganTStevenson @Ljt019117161 @johnjhorton The problem is that earlier tools had much lower accuracy narrating a complex doc like a research paper PDF. Paper2Audio accurately narrates your docs using high quality voices, building upon recent AI progress in multiple areas. Happy to discuss or answer more questions.