Elfin2

@elfin2

96 Following

2 Followers

Elfin2 pfp

Begin again. Be Rewarded. Keep growing.

0 reply

0 recast

0 reaction

Elfin2 pfp

lol personally I think he should keep it and spoil himself It was earned fair and square

0 reply

0 recast

0 reaction

ash pfp

everyone bidding on Dan’s posts this is supposed to be about supporting creators lol

11 replies

2 recasts

50 reactions

Elfin2 pfp

If you miss the sunrise 🌅, don’t be sad—enjoy the stunning sunset 🌇

0 reply

0 recast

0 reaction

Rakshita Philip pfp

Rakshita Philip

naah i think you’re pretty awesome for doing that

1 reply

0 recast

1 reaction

Elfin2 pfp

Just another day in Ireland where the sheep have better real estate than I do! 🏞️🐑 Grazing lakeside in the stunning Gap of Dunloe—rent-free, stress-free, and with million-dollar views. ⛰️✨

0 reply

0 recast

0 reaction

Elfin2 pfp

This is the right answer.

0 reply

0 recast

0 reaction

ciniz pfp

quote this post with your best anime recommendation. only top-tier picks i'll add good ones to my list and repost if i've watched them tipping everyone

17 replies

29 recasts

130 reactions

Elfin2 pfp

0 reply

0 recast

0 reaction

avi pfp

enjoying reading these agent architecture blog posts recently in summary how claude deep research product work is as follows. there is a lead agent powered by opus 4. it accepts user's query and creates a plan. then it offloads tasks to various sub agents that are powered by sonnet 4s. sub agents mostly run in parallel and provide separation of concerns. sub agents themselves are capable of planning too. they often revise their plans and change trajectories if needed. tool usage and planning are interleaved during the whole process. secret sauce of a multi agent research tool is compression. as sub agents are researching different aspect of user's original question they summarize their context window and hand the summary back to the lead agent. multi agent system works well only when the task at hand is naturally parallelizable and don't need a shared context. most tasks are not. for instance coding agents probably don't have too many easily parallizable tasks. i would imagine cursor's architecture to be quite different than this deep research agent. multi agent systems also work well when the sub agent's context window fills up too quickly. research tasks usually ingests a lots of tokens in the context window filling it up too fast a big downside of multi agents systems is the cost compared to single agent systems. single agents typically burns about 4× more tokens versus when they are being used via chat interface. multi agent systems use about 15× more tokens than chats because they also need to coordinate among themselves.

2 replies

3 recasts

22 reactions

Elfin2 pfp

Enough to save me

0 reply

0 recast

0 reaction

Elfin2 pfp

Congrats homie well deserved

0 reply

0 recast

0 reaction

arob.base.eth pfp

Insane week on the purple app, I’ve never been in the top 200 before and have a chance to make the top 180 if i don’t drop too much overnight. Plus I’ll be visiting Coinbase tomorrow to give my opinion on some new tek. Based 🔵

26 replies

7 recasts

74 reactions

Elfin2 pfp

0 reply

0 recast

0 reaction

Joseph Maggio pfp

Uncle @dawufi Why do they call you dawufi ?

3 replies

0 recast

11 reactions

Elfin2 pfp

Stockholm, Sweden 🇸🇪

0 reply

0 recast

0 reaction

Rakshita Philip pfp

Rakshita Philip

it’s a good day to be $brunette

5 replies

2 recasts

13 reactions

Elfin2 pfp

0 reply

0 recast

0 reaction

Elfin2 pfp

would love to check it out and always down to chat photography in the app!

0 reply

0 recast

0 reaction

aneri pfp

@aneri.base.eth

Starting tomorrow, I’ll be sharing more Coinbase Wallet beta invite codes with those posting from the beta often. Active users like you are best positioned to bring in others who can help make the app even better.

12 replies

5 recasts

95 reactions