Sagar

I love my country and it’s people but sorry an IPO doesn’t show ground reality I am travelling by train after a long long time it’s not a pleasant sight the video shows condition of a local train in not so important place of West Bengal at 6:40 pm Now, imagine Mumbai or any other major city and how many people are travelling like this btw I’m in an E2 and the condition of it is a topic for another day

FDE interview question on context engineering What if a weaker LLM agent writes misaligned goals to the memory and a strong one executes it How do you prevent such scenario?

‼️This paper can vaporise 100s of startups overnight soon OpenAI, Anthropic and other labs will implement this technique then all context management layer where Mem0, SuperMemory, Inherent etc. sits will be vaporised For a forward deployed engineer context engg is the moat make sure you stay updated with these papers

AI engineers especially forward deployed folks who work with parsing documents read this paper building the first 80% is easy but the UAT sucks as accuracy fumbles this can help you get your system over the line

I have a daily newsletter that keeps me updated on Context Engineering It summarises recent papers extracts the key messages all in TLDR style and written like a MBB consultant (like me) you should build yours too or DM your email if you want to be part of it.

Good Sunday afternoon read

Noob: bro just use an agent Pro: why do we need an agent? here is a simple matrix to answer the pro

this is the biggest unlock for me by @claudeai

Did anyone implement this?