@kanikairl same product but better implementation, cheaper
and which people can actually get access to and use in their daily life.
IMO these are the things which makes a difference.
@tanujDE3180 No, Its not a RAG
R in RAG stands for retrieval
so a RAG always searches for relevant chunks from the documents and feed those to LLM to form an answer.
While in this case, the whole document has been fed to the LLM in a prompt. There's no retrieval happening.
@riddhi_builds Yesterday at a party I met a guy, he charges 2 lakhs per month to make reels for brands.
He takes photo and audio of owner and make deepfake reels using AI and promote their brand
@im_Abhi10600 Hmm I agree with this point
Recently when JEV was released, typesafe told that a different kind of post trianing step is used to train it called RLCD, but we don't know what RLCD actually does.
@im_Abhi10600 Yeah I agree, but if models are open sourced then who will pay for gpu?
They use a hell lot of compute.
And if we think of running models locally then everyone can't afford it too.
@shivambhadani_ I am a 3rd year student, suggest me what should I do to make my skills better in Agentic AI, Harness Engineering, Context Engineering etc.?
Amazon recruited 50 people from my college for SDE(6m intern + performance based PPO)
Anyome know what's amazon is upto?
Why are they hiring so aggresively from past few months?
@varshaa_twts If you will use coding agents for devops then
For a task that can be done in 4 steps, AI will tell to do in 10 steps, so always first understand the flow the implement