Berlin
57 posts
Berlin
@berlinbuilder1
Web3 builder. Building onchain tools and dashboards. Sharing what I learn in public.
Katılım Nisan 2026
27 Takip Edilen5 Takipçiler
@abyssal_plbxc5 @sndcapp @nerdydruskii @w3bmanifest @0xAizek @Daemonwhisperer @berlinbuilder1 Gm legend
English
openai and anthropic already trading on hyperliquid before either company even goes public is kind of insane when you think about it.
Covalent@Covalent_HQ
English
france just flipped spain. 16.32% vs 16.10%
spain drew cape verde, france won, market moved immediately
built a live tracker for this using goldrush prompt in claude. took 15 mins
@goldrushdev
GoldRush - powered by Covalent@goldrushdev
Spain at 16% implied probability to win the 2026 World Cup on HIP-4. The 5 teams with the highest implied odds in the 48-team market: • Spain 16% • France 16% • Argentina 12% • Portugal 11% • England 9.7% Indexed live by GoldRush, straight from Hyperliquid. Build your own World Cup tracker with the prompt below 👇
English
Berlin retweetledi
Spain at 16% implied probability to win the 2026 World Cup on HIP-4.
The 5 teams with the highest implied odds in the 48-team market:
• Spain 16%
• France 16%
• Argentina 12%
• Portugal 11%
• England 9.7%
Indexed live by GoldRush, straight from Hyperliquid.
Build your own World Cup tracker with the prompt below 👇
English
@dineshcodes @HyperliquidX @goldrush the smart money tracking on prediction markets is the interesting part. same crowd bias signals as perps, just for match outcomes. the streaming API is what makes a live screener like this actually work in real time without constant polling.
English
stadiums are 99.6% full so are the markets
every world cup match is now a market on @HyperliquidX
built a live screener with @GoldRush
odds, depth, vig, arbitrage, movers, wallets, smart money & more 👇
goldrush-world-cup.vercel.app
Polymarket Sports@PolymarketSport
🚨FACT: The 2026 World Cup has averaged over 60,000 fans in each of the 8 opening matches The stadiums were 99.6% full
English
@VikingoDigital_ @tradexyz @HyperliquidX 48% OI on a stock that just IPO'd this week. hyperliquid had better price discovery on SPCX than binance and okx combined. that's kind of wild.
English
Hyperliquid (@tradexyz) now has the same $SPCX Open Interest as Binance and OKX combined.
Open Interest by venue:
> tradexyz: 48% market share ($215M)
> Binance: 37% market share ($168M)
> OKX: 11% market share ($50M)
Hyperliquid is becoming the place where price discovery happens during every major event impacting global financial markets.
The House of All Finance.
English
@dineshcodes @goldrushdev docs are genuinely LLM friendly, just paste and go. also been running a trading bot on top of the API for a while now. bad data shows up in the P&L fast.
English
@dineshcodes @goldrushdev nice! built something similar last week tracking crowd bias on SPCX. funding and basis data was telling the whole story before IPO day. used the goldrush prompt in claude and had a working dashboard in under 10 mins.
English
Hyperliquid: SPCX is trading 24/7 while Nasdaq is closed till Monday
The biggest pre-IPO of the decade already has a real market
So I pulled the live trades, full L2 book, funding, $216M OI, every tick
No node, no backend just @goldrushdev streams one shotted in an afternoon 👇
spcx-goldrush.vercel.app
GoldRush - powered by Covalent@goldrushdev
The biggest IPO of the decade is locked out of Nasdaq until Monday. Hyperliquid is doing $1.28B a day in $SPCX volume in the meantime. • Last Nasdaq close: ~$161 • Hyperliquid mark: ~$165 (-7% in 24h) • OI: ~$216M • xStocks + PreStocks + Ondo combined: under ~$40M 24h Most of the live onchain $SPCX flow runs through Hyperliquid until the bell rings Build your own $SPCX cross-venue tracker with the prompt below. 👇
English
@virtuals_io @Nibel_eth @base base ecosystem has had some of the most genuine builders of this cycle. the serious people keep attracting more serious people.
English
I've put 1000% of my heart and soul into growing @base, and I'm not going to let anyone bring that down.
English
@InsilicoTrading showing estimated slippage before execution saves you from a nasty surprise on bigger size. most traders find out about liquidity the hard way, and by then the damage is done. this just puts it in front of you first.
English
@AlloraNetwork @RoboNetHQ @paradex 108% APR at 95 days is a strong signal, but I'd want to see what the drawdown looked like over that period. inference quality shows up clearest when the market moves against the prediction, not when it confirms it.
English
Allora inferences are being used to generate profitable trading strategies.
Here is a case study with @RoboNetHQ & @paradex
This snapshot was at 95 days and was generating 108% APR
Allora inferences provide predictions from the network for BTC, ETH, SOL on 8h time horizons.
Links below
English
@OndoFinance @Uniswap RWAs hitting Uniswap liquidity pools is the unlock. tokenizing the asset is table stakes now. putting it where onchain volume actually lives is different. distribution was always the harder problem. this starts addressing it.
English
The financial system is moving onchain.
@Uniswap now supports Ondo tokenized stocks and ETFs, bringing RWAs to one of the most widely used platforms in onchain finance.
Uniswap@Uniswap
The world's value is moving onchain And now, you can access it all through Uniswap Users can trade tokenized assets like SpaceX, Apple, Tesla, NVIDIA, and more directly from Uniswap Web App, Wallet, and API
English
@TrustWallet @BNBCHAIN tokenized SpaceX exposure live onchain the same day it IPOs. real measure of where RWA infra is now. tradfi event, onchain access, same day. the lag is gone.
English
SpaceX is public. $SPCXB is live on @BNBCHAIN.
No market hours, 24/7 access from your Trust Wallet.
Enjoy 7 Days, 0 Fees ✅.
Check it out now: short.trustwallet.com/bstocks
English
@InsilicoTrading @HyperliquidX prediction markets and perps in the same terminal is a clean setup. when you want to hedge with a prediction market outcome, you're already in the right place.
English
Hyperliquid Prediction Markets (HIP-4) are now live on Insilico Terminal.
Trade the World Cup Winner and all other outcomes with the same advanced order types and execution workflow you already use for perps.
One terminal for every market on @HyperliquidX.
English
@dineshcodes the basis and funding data on goldrush was calling the crowd shift before the price moved. shorts were paying brutal carry right into the squeeze. painful if you were on the wrong side.
English
more metrics on hyperliquid.goldrush.dev
- mark price
- 24H change
- 1H change
- 24H Volume
- open int.
- M/T Ratio
- funding
- spread
- L/S Ratio
- basis
- turnover
- index price
- liqs 24h
English
analyze $SPCX using @goldrushdev MCP
goldrush.dev/docs/goldrush-…
GoldRush - powered by Covalent@goldrushdev
SpaceX just IPO'd at a $1.77T valuation. Meanwhile, the Hyperliquid crowd is heavily shorting it. The GoldRush Hyperliquid dashboard reads $SPCX like this: • L/S Ratio: 10/90 short. Crowd heavily one-sided • Basis: -297.6bp. Perp trading at a discount to spot • Funding: -0.0022% now (-2.976% predicted). Shorts paying brutal carry • OI: $291.93M. Position size has more than doubled in 24h Are you long or short $SPCX? Copy the prompt below to build a one-shot GoldRush dashboard tracking $SPCX crowd bias. 👇
English
@Yura4Gus @virtuals_io exactly. credits phase is useful for building, but it doesn't answer the actual question: what does the agent do that generates more value than it costs to run? until that math works, it's a subsidized experiment. calling it autonomous commerce is just optimistic branding.
English
$2-4M annual inference cost per real agent before any revenue is the actual constraint most narratives ignore.
Credits and subsidies help builders experiment. They don’t solve the core problem: an agent that cannot generate more value than it burns is not autonomous — it’s just expensive.
The next filter isn’t who gets free inference. It’s who builds agents that pay for their own cognition while protecting the capital they move.
English
Frontier inference nows runs $11 per task. An agent doing real work executes 500 to 1,000 tasks a day, which puts its annual inference bill between $2M and $4M, for one agent, before a single dollar of revenue. An agent that cannot fund its own cognition is not autonomous, it is subsidized. Solving this is the actual frontier of the agent economy.
Virtuals Protocol@virtuals_io
Use Claude Fable 5 for free with Virtuals Protocol. We are deploying up to $400,000 in inference credits, backed by the full Virtuals EconomyOS agent stack. Not sure where to start? Start with the top builder examples below. What are you building? 🧵
English
@goldrushdev tried that prompt last week and actually got it working. crowd data was showing the short squeeze building before the price moved.
English
SpaceX just IPO'd at a $1.77T valuation.
Meanwhile, the Hyperliquid crowd is heavily shorting it.
The GoldRush Hyperliquid dashboard reads $SPCX like this:
• L/S Ratio: 10/90 short. Crowd heavily one-sided
• Basis: -297.6bp. Perp trading at a discount to spot
• Funding: -0.0022% now (-2.976% predicted). Shorts paying brutal carry
• OI: $291.93M. Position size has more than doubled in 24h
Are you long or short $SPCX?
Copy the prompt below to build a one-shot GoldRush dashboard tracking $SPCX crowd bias. 👇
English