AI that can read 150,000 words at once

This video is not sponsored by Anthropic and OpenAI, just in case if someone is wondering LOL

Check out my newsletter here:

Claude 2.1

Lost In The Middle Paper

Greg Kamradt's Findings

This video is supported by the kind Patrons & YouTube Members:
🙏Andrew Lescelius, alex j, Chris LeDoux, Alex Maurice, Miguilim, Deagan, FiFaŁ, Daddy Wen, Tony Jimenez, Panther Modern, Jake Disco, Demilson Quintao, Shuhong Chen, Hongbo Men, happi nyuu nyaa, Carol Lo, Mose Sakashita, Miguel, Bandera, Gennaro Schiano, gunwoo, Ravid Freedman, Mert Seftali, Mrityunjay, Richárd Nagyfi, Timo Steiner, Henrik G Sundt, projectAnthony, Brigham Hall, Kyle Hudson, Kalila, Jef Come, Jvari Williams, Tien Tien, BIll Mangrum, owned, Janne Kytölä, SO, Richárd Nagyfi

[Video Editor] @askejm

[artiIIerycrusader]

Easy to decrease hallucinations by simply not telling us anything at all.

[SPIKEASAURUS]

Claude Debussy still my favorite Claude.

[avi]

I like the presentation style and pacing of this video. The editing is also pretty good, for example it feels like the video elements and animations move you along / push you forward as a viewer (e.g the ChatGPT logo fixed over a Claude thread being scrolled). The video length probably could and should be longer. Perhaps it could be augmented with the demo of a relevant experiment to fill in the missing run time and increase the overall value. The swoosh effect on transition is a bit much, would maybe tone that down a bit.

[camwha]

Awe, longcat! That’s a meme I haven’t seen in ages!

[pareak]

I feel like most people really don't have any understanding on how useful large context window really is and how huge this problem is in the LLM we have at the moment. I guess that's also why Anthropic didn't make a big deal out of the 200k context window.

[anotherslowtown]

can someone explain to me what this means for RPing with AI? will it be better at remembering the RPs? (nevermind.)

[brekol]

i have to try claude again, last time it kinda dissapointed me..

[l]

Imagine how fast AI can do your homework in the future 😅

[ernesto.iglesias]

You can't compare them with these graph because GPT-4's one have a different scale hiding most of it errors

[gabrielv.]

I want free AI image or video generations, specially videos... Any tips? (no login if possible, if i have to login, i want unlimited acess)

[kitchenpotsnpans]

The solution is analog computing, i already knew this mauhaahhahaha

[Gustavoooooooo]

also Claude's not available in 100 out of 195 countries

[fnytnqsladcgqlefzcqxlzlcgj]

Lol it can fit the bible twice, or like 1/8th of Srimad Bhagavatam or 1/20th of Mahabharata lmao

[antonberkbigler]

The fuck you mean it’d take a human 130 hours to read just 150k words? I’ve read twice that in half as long multiple times, and that’s including things like meals and sleep. Decreased sleep usually, but still. I have tech addiction.

[djtyros]

First comment by a real person. Yay, im alive! SWITCH ME OFF