ykumards

IMO it would be hard to reproduce the results using autoresearch setup.

To get CLIP to work properly we typically need large batch sizes. So the experiments in the original paper were quite heavy, and ran parallel across 8 GPUs.

Sorry I could have worded this part better.

The docker container didn’t have network access. Claude didn’t have permission to execute anything other than the run.sh bash script, which would orchestrate the docker run

That’s true! It felt a bit flippant to give medical data to an agent. Also, I wanted to see if the model would work in other domains!

Thank you, glad you liked it!

I'm subscribed to Huggingface's Daily Papers newsletter but going through each paper individually was a chore.

So I created a GH workflow to pull the papers, use an LLM and TTS to generate the podcast daily.

Hope you like it!

True :)

https://xkcd.com/303/

Thank you so much! :)

I’ll definitely try to add audio. Local TTS models are getting better and I guess I can feed the headline to create a podcast-like feed

That’s a great point! I could add some “roles” that people can choose for the kind of papers they’d like to see.

That was exactly my reaction after reading the first few paragraphs. “How did Eve circumvent the email verification step?”

Netflix should totally fix this, even my local public library asks for email verification after signing up.

That's totally fine, as long as you are aware of the problem and decide that it doesn't matter to you. But they have the moral(legal?) obligation to let us know how they're using our data.

I share your concerns about losing Facebook data of the past, but frankly you didn't address any of the concerns that were listed in the article.

The issue is what they do with our data and whether you'd be comfortable letting them use your data this way. It's also due to their shady practices like not disclosing changes to privacy policy. This is especially worse with non-technical people like my mom using their platform. Who knows what kind of bs they're peddling onto her wall.

I think it's awesome that people are finding ways to get off Facebook. It's like getting out of a heroine addiction or something:)

* uBlock Origin - Lightweight adblocker

Link: https://chrome.google.com/webstore/detail/ublock-origin/cjpa...

* Tab Snooze - Removes clutter by hiding a tab for later. We can pick the time & date. Can use this as a reminder.

Link: https://chrome.google.com/webstore/detail/tab-snooze/pdiebia...

* One Tab - Push all current tabs into one tab. I use this with Tab Snooze, mostly. All tabs related to a project are snoozed till I get to work on them again.

Link: https://chrome.google.com/webstore/detail/onetab/chphlpgkkbo...

* Papier/Workflowy New Tab - Have a text document in the new tab page. I usually have to-dos or jot down some quick notes here.

Link1: https://chrome.google.com/webstore/detail/papier/hhjeaokafpl...

Link2 : https://chrome.google.com/webstore/detail/workflowy-new-tab/...

If you have a workflowy account, this syncs well and is similar to Papier.

https://chrome.google.com/webstore/detail/workflowy-new-tab/...