this has everything i need to build a wallet, a terminal, a trading bot, a launchpad

first 10k token launched on intear launch, $NEAR6. Congrats btw @NEARProtocol

the first runner on intear launch must happen this week, at any cost, or i stop the launchpad. everyone whose launch goes over 10k mcap (not botted / wash traded) will get a share of $tear from me

This better be big, Eight

New tasks in $8888 airdrop: Supply liquidity on Intear DEX for NEAR-USDC pool. It also has 12.5% APY btw. Making the best DEX is one of my goals for next year

Just learned that you can deposit tokens from @shitzuonnear bridge straight into @near_intents confidential deposit addr. Sounds like the most useless knowledge in crypto,bridge into another bridge, but was really useful for me since near com doesn't support xaut deposits on near

Number of quantum-safe accounts on mainnet has gone up 5x since intear wallet added post-quantum support, number of daily active post-quantum accounts and transactions went up around 10x. Still not a big number though. wen near intents support

Meteor and Intear support post-quantum cryptography now. Be among the first to migrate! Made a little dashboard for migration stats: arewepostquantumyet.com

Building 8888 confidential markets on near btw

just added ml-dsa on @intea_rs wallet, currently in import-only mode. share screenshots of importing quantum-safe wallets into intear wallet!

llms are so good at optimizing code (and other tasks where 1000 bad attempts can be better than 1 good attempt). i optimized my "simulate the entire game with 20k units with independent ai on infinite map" benchmark of my upcoming rts game from 150 seconds down to 0.6s. took me a day, without llms would've taken like 3 days

currently building an RTS game, my favorite genre, loved it since first time playing warcraft 3, one of the first games i've ever played. looks like a clash royale ai ripoff but im building an engine first, and need something to compare to. after that, i will probably build a language learning app / course / social platform, because im so tired of people telling me that humans can't reach fluency in less than 3 years, or apps giving me options from 5 to 20 minutes a day, teaching 5 words a day to maximize subscription retention. once you have one or a few basic survival skills to get a real job (coding, art, etc.), languages are the most useful thing that unlock a ton of content otherwise not accessible. takes time but these DLCs are worth it + learning is fun (at least, for me)

i chose the right horse in the $50 million race

bullish for intear dex, now its the safest dex on near

what music player do you use? looking for a new one. the one i was previously using is written in C and i'm scared to run it (i discovered it by seeing some strings truncated in UI, that usually doesn't happen unless there is some memory bug lol. checked github and it's indeed C)

6 days in, ~8 hours of coding, progress: - basic tokens (no mint / freeze / burn authorities yet and no metadata, metadata will be added today) - basic smart contract runtime (production-grade coming later today) - named accounts, subaccounts - multiple keys (ed25519 for now, p256 and secp256k1 coming today or tomorrow) not done yet: - gas costs - relayer tech - nft - libraries for development - consensus mechanism with multiple nodes - rpc - archival rpc experimental ideas: - fast shard (for orderbook, more centralized but faster block times) - on-chain orderbook & amm - built-in price oracle that takes average data from orderbook and amm from the last X minutes - geo shards (for minimal latency) - thresholded signature network - slow shard (higher gas limit for long executions, such as nested wasm runtimes or complex zk proof verification, maybe some on-chain ai, maybe a docker sandbox for each transaction to achieve native speed, idk) - ai agent oracle (describe a harness and all the prompts and tools, each validator will run a sandboxed ai agent, and 67% of agents must agree on the outcome, otherwise a fee is charged and error returned. maybe good for things like prediction market resolutions, where the answer is yes/no or a number) - http oracle (similar to ai agent oracle, validators must agree on outcome, so only useful with static content) - bridge (between shards and between near<>tear, using yield/resume and light client proofs, so things like ft_transfer_call or function calls can be forwarded back and forth transparently like a normal cross-contract call)

my blockchain will likely have the best feature-to-code ratio on the planet. insane to think about in the times of ai, but that's my goal. every line of code is attack surface, and with ai getting better at finding exploits every month, there's no good way out of the loop other than reducing attack surface. by "the loop" i mean pointing ai at every file of tour code 100 times every time a new model is released. people actually do this. but a hacker with money / ai credits will point the model at your code 2000 times (real number from some anthropic post or research) and find something. kiss 😘