idk how many of you have paid attention but this is how chainflip's april looked: > $72.4M swapped in the first week alone (avg swap size up 33.2%) > by week two, lending TVL hit $1.43M apr 14 was the biggest swap day in 70 days ($30.7M) > in week three, all-time swap volume crossed $7B > in week four, FLIP burn crossed 9,000,000 all time (~$1.82M). while the minimum active bid doubled to 204k FLIP. multiple integrations kept landing (@THORSwap, @leodexio, @safepal @SwapSpaceCo, @nornlink, @AgentChud etc) as april came to an end, chainflip marked $3.13M in a single swap and more than $2.5M lent in total and just before the month wrapped, lending 2.0 proposed: supply as collateral, btc earns Boost seems $FLIP is winning more details ⬇️:

what I am doing to optimise my content with AI i started building a personal AI knowledge layer from scratch. using my X archive, my bookmarks, Claude, and Obsidian. here's exactly what I did and why: every time we use AI, we explain the same things repeatedly. from research history to goals and the tone of posts. a month ago, I came across Shann’s AI marketing articles. I read a few and practised them which really helped me get active here again but i also bookmarked most of them i wanted to follow everything yet i kept procrastinating. 3 days ago, i saw his AI knowledge article. with detailed practical steps to replicate his exact setup for myself. it has 2 layers to make your agent smarter. 1. a dynamic knowledge base that keeps growing with every use (your notes, tweets, articles - processed into a linked wiki) 2. a static brand foundation that adapts your voice and rules - never changes) i asked claude to study the article and help me build It took me less than an hour. > step 1: requested my X archive (took me 2 days) i uploaded the zips to Claude and asked it to read the full archive and extract > step 2: claude built the wiki structure. folders for concepts, entities, sources, outputs, SOPs. then it generated: - entity pages (projects i wrote most about) - concept pages (defi, stablecoins, GEO/SEO, copywriting, marketing) - a brand-foundation.md with my actual voice patterns pulled from the archive this really was not generic, btw this post is also written by claude (~60%) in my voice which is claude said, and i quote "direct, process-transparent, and no hype" > step 3: I opened it in obsidian vault. the graph view already had connections. For example, GEO/SEO is linked to marketing, linked to copywriting. a visual map of all the concepts i’ve written until now. > step 4: i connected Claude Code to the wiki folder. dropped a test note(copied the text of a bookmark i had) into raw/clippings/. ran one command. it classified the note, extracted new techniques, updated my copywriting page, flagged entity stubs, and logged everything. what you need to build this: → your X archive (settings → request your data) → obsidian (free) → claude claude code → @shannholmberg's article (will link in replies) (i also used chatgpt to verify every step of mine so i don’t run into roadblocks didn’t use claude because i didn’t want to exhaust my tokens :P) if you're a researcher or a marketer/content writer - this is worth building.