sabareesh

10 ms·

91.
by sabareesh 2y ago
Nailed it, More often it is the water than the other ingredient even with the coffee
92.
by sabareesh 2y ago
Well it is time to finally learn the lesson here. We dont want companies self-auditing, self-inspect, self-govern themselves
93.
by sabareesh 2y ago
These are not apples to apple comparison, as this is running across GPU and much bigger model
94.
by sabareesh 2y ago
This much bigger model (500M), P2P is enabled via Mailbox. It is expected because of memory to compute ratio
95.
by sabareesh 2y ago
Well here is a comment on 4090 https://github.com/karpathy/llm.c/discussions/481#discussion...
96.
by sabareesh 3y ago
Did you say first manned landing ?
97.
by sabareesh 3y ago
Please think again what is the point of going back to moon if we cannot establish a base and stay longer ? It is not about just going to moon we want to stay there and establish a base
98.
by sabareesh 3y ago
I believe again we are shortsighted here, it is not about just going back to moon but it is to establish as base. Based on wikipedia starship can take 100 tons to Moon after several refueling, which makes sense. We have already had our boot
99.
by sabareesh 3y ago
Can you please explain why trashing starship makes sense.
100.
by sabareesh 3y ago
Destin's video was shortsighted but I believe his new video is better on understanding it is not about just flags any more.
101.
by sabareesh 3y ago
Even if you pay it is hard to get such a high quality content!
102.
Trekking Through Retirement: Tax Strategies in Retirement [video](youtube.com)
1 points by sabareesh 3y ago|0 comments
103.
Tax Strategies in Retirement(3nickels.com)
1 points by sabareesh 3y ago|0 comments
104.
by sabareesh 3y ago
GPT4 Turbo is top of the class it does RAG very well, it is important to provide good context with help of Vector DB but if you cannot provide relevant document it cannot do much. All the opensource models are super bad at this and mostly i
105.
by sabareesh 3y ago
I want to see benchmarks for RAG. Most of the models are not very good with RAG
106.
by sabareesh 3y ago
If we can generate Methane for example from CO2 in the atmosphere this could potentially solve this problem
107.
by sabareesh 3y ago
We are changing how things are being done. Before NASA was responsible for everything and more hands on approach but now it is hands off approach with fixed cost contract. Because of this you can expect things to work the same way as before
108.
by sabareesh 3y ago
Loved the most of the presentation but i felt he lacked the big picture, it is no longer about just going to the moon but this time we want to potentially settle and expand to other planets. Because of this we are not just trying to replica
109.
by sabareesh 3y ago
Does anyone know how to load this into ollama
110.
by sabareesh 3y ago
~ 34 million embeddings , Seems this is a good enough sample for our use case.
111.
by sabareesh 3y ago
Well tried it with ~ 34 million records , which is good enough sample for our usecase
112.
by sabareesh 3y ago
Other day i was trying out Azure SQL for storing vector DB. After a POC i was able to get results on sub 1 second to get the results with 2 core serverless instance. It begs me to reconsider does dedicated Vector db really worth it ? Here
113.
by sabareesh 3y ago
Well definitely reduced times i use google as a developer
114.
by sabareesh 3y ago
Care to explain why. I am in favor of all the maturity of several libraries and framework
115.
by sabareesh 3y ago
I am noticing there is enough entropy on the agent replys and also as we can adjust the temperature they can work together to provide a better answer by peer reviewing.
116.
by sabareesh 3y ago
Multi Agent has several potential, I am having more confidence as there is some level of entropy on agent reply that makes it a worthwhile
117.
by sabareesh 3y ago
well basically this is what i have been doing since functions introduced part of chatgpt api
118.
by sabareesh 3y ago
If anything this is making it safe for the owners because pretty much all the third part apps have full access to vehicle because some owners shared their password to some random third party company so that they can have some additional fea
119.
by sabareesh 3y ago
Found this lawn mover using similar strategy https://www.ecoflow.com/us/blade-robotic-lawn-mower
120.
by sabareesh 3y ago
MultiAgent could potential solve this easily

More ›