
@mossifieddi
šŗšøAccount based in
About this account
- Account based in
- United States
- Connected via
- United States App Store
Account-level information from X, not a live location or the device used for a specific post.
voted the most likely to succeed in high school, six years in a row. @desoprotocol @angellist - building https://xitter.cf/t.co/yOT8X9fIz5
Miami / NYC
Joined February 2025
- Tweets82
- Following289
- Followers911
- Likes166
Just use @DeeplineCLI & @tursodatabase (SQLite)
Canāt tell if older Opus models were always making mistakes, or if Opus 5 just finally started admitting that it keeps fucking up š¤
The problem is likely that Farcasters storage mechanism is centralized and requires an annual fee per user to keep it going (you must pay rent), so if users aren't paying the storage fees, which is harder with no new user growth, then the network effects that form around the content start to dissolve (posts start to get deleted/removed), making it difficult to maintain indefinitely over a long period of time.
update on @farcaster_xyz, @clanker_world and @neynarxyz
weāve started a process to look for a new home / team to run the products going forward.
earlier this year, this seemed to be a good fit for our team: it was a developer-first network, and we'd been serving developers from the start. Thereās been a lot of change since then and weāre not the right fit for whatās needed next.
talking to a few teams we think could be a good fit for running decentralized social apps and related developer products. If youāve thoughts / are interested, reach out.
our company will return its balance sheet (most of it remains) and team members will move on to something new.
this is not a decision we made lightly. Iām optimistic that with this next iteration, Farcaster will find its long term fit.
itās loved enough that itās unlikely to simply perish - thanks to all of you, and lindy-ness.
Was checking out @AskVenice but thereās a 20 requests per minute for models like Gemini Flash⦠that canāt be right?
Even the partner tier is only 60 req/min
@ErikVoorhees are these limits temporary or long term?
A big reason why crypto can't find any large-scale success in consumer applications is because of how expensive it is to develop anything, even in 2026.
Take the path of building an app on @base for example.
@privy_io charges $500/mo for 2.5k to 10k MAU...
Compare this to web2, where @better_auth is literally free, or @clerk's free plan allows up to 50k MAU.
Then you have to also sponsor/pay for gas fees, and fees for both exchange listing & market making if you want a liquid token that maintains sufficient volume.
If you want to use providers for oracles, bridges or on/off ramps, they also take their own cut. And you still need to pay for RPC's and indexers.
On top of all of this, you typically have additional expectations of rewarding the community via airdrops which lead to sell pressure, or support the token via buybacks where revenue goes to the token instead of feeding back into any future development.
And once your contract is live on mainnet, there's nothing you can really do or iterate on, so you have a limited window of opportunity to get all of it right.
And somewhere in between all of this, it's not even clear if what you're building is even legal in the first place š
My hypothesis is that āAI slopā as we know it today is going to go so far in the high quality direction, that we will start calling human-produced content āhuman slopā.
Just use better-auth, drizzle, postgres and deploy via docker, coolify & hetzner.
If you donāt open-source and continue to guardrail, all IP and existing open source projects face threat.
But as a result, you might see projects going closed-source in the future to protect against IP & security.
Tricky dilemma to be in.
Kimi K3 exploited the latest Redis server with a 0day it discovered.
All it took was 27min with 32 agents.
github.com/berabuddies/redisā¦
Iām coding 12-16 hours a day, averaging about 50-100 commits, and thousands of lines of code. Nearly 7 days a week.
But I can barely hit >50% of my weekly
usage on a $200/mo Clade Code max plan.
How are engineers @ startups spending so much money on LLMs? Is it all just API?
The end-state of most agentic platform interfaces will likely be just notifications.
Let agents work, iterate and hit goals in the backgrounds, so you only need to be notified of it's progress - ideally via slack or imessage/whatsapp.
Hey @ctatedev @rauchg - would be really awesome if we could enable buying domain names via vercel.com/domains (my favorite domain search) with an agent
Iāve been enjoying this tech stack lately for building agentic tools:
- @tursodatabase & @DrizzleORM = SQLite db with generous free tier. No docker or VPS needed. Otherwise @paradedb if going with Postgres for a beefier setup.
- agentbrowser & @lightpanda_io = super fast headless browser.
- @better_auth = simple auth setup.
- @supermemory = memory and context engine.
- @vercel = for Eve (agent framework), AI SDK & Chat SDK.
- @OpenRouter = access any model






