0x008

Great production value

RAG is THE rage right now. Everybody is talking about it in enterprise world because they want to make all their legacy documents searchable.

I think if you run a cluster in a trusted environment it should be more efficient to use ray or something similar

is the $2K you mentioned the total cost of ownership?

<Wow i never knew their music was so complex. That video is amazing

For me it is a interesting consequence of our modern ways of communication which is characterized by being always-online and always expecting immediate response that people turn to giving the new name of "flow-state" to what is essentially just "uninterrupted work".

yeah it's an easy way to update firmware on microcontrollers

my excuse is that my company VPN is tied to chrome.

usually "architects" are just people there to keep the communication rolling and make sure devs or teams don't work in silos all their life.

probably your communication skills were too good for management to passt up the opportunity of promoting you.

wonder how many applications they get through the application form.

> If your pavement has potholes, just learn to jump over them.

And if you have nested list comprehensions this totally stops making sense somehow.

The prices there are totally not real. The A100 x 8 is listed for $5, whereas amazon price calculator lists it for $32 per hour...

In times like these, when there is so much opportunity and so many use cases are easily recognizable due to the AI hype and so much information going around, anyone can be a product manager. When the tides turn and we enter the next AI winter, lots of these product managers will move on or be cut.

It's very easy to be a product manager in times like these, it's not easy being one where opportunity is rare and hard to find.

you can use pyenv to have multiple parallel python versions installed and you can choose per virtual environment which version you want to use.

Noticed this also today.

Any information on why this is needed? Do I again need to migrate all my bookmarks?

Any tips for selfhosted alternatives with good extensions?

edit: here is the justification from the author:

> Previously this permission have been opt-in to unlock features like "Save all tabs" and "Saved page indicator". But those features are esential and should be available right away, isn't? That's why I decided to require tabs permission by default. I have received dosens of emails from users complain about not working saved page indicator.

https://help.raindrop.io/changelog/#560 ... still don't feel too comfortable with that. permissions like this should be opt-in. just stating it "should be default" sounds strange to me.

NVLink was discontinued with 40 series nvidia consumer GPUs.

They are, but data center GPU (ampere and above) are twice as good in performance per watt.

Can somebody ELI5 Rasa for me? I read through the README and I still don't get what it does.

1password has a great UI imo. and they now support ssh keys as well (albeit a bit strangely, but at least they do).

> HyperSVMRanking (Such Vector, Much Ranking) algorithm proposed by Andrej Karpathy (2023) https://arxiv.org/abs/2303.18231

The API access is not a subscription, it’s pay as you go.

Also note if you pay by token for GPT3.5 it is (in my experience at least) a lot cheaper than the subscription ($0.64/day but also when you not use it) vs. ~0.20€/day but only if I use it. I use pay as you go and it comes around as a few $ per month versus the flat $20.

Not at all. Where are you from?

Oh yes but we are restricted, very much so, by the assumptions what other people might think of us or how we could be exiled from society if we say x or y forbidden thing.

Much like GPT fine tuned to not say certain things so are our brains by the parenting we receive and reactions we experience in early life.

It’s not only interesting but also necessary. What is language if not a compressed version of all human culture?

I think it’s called davinci

the standard GPT3 version used in chatgpt is gpt3.5-turbo. it is a pruned version of GPT3 (trimmed down), to have faster response times and less computational cost.

It is not generally available though?

How are you using GPT-4 if there is no API for it currently?