Finding the next Web3 gems before they explode. I'm here for the tech, the cute utility.
United Kingdom
Joined June 2020
- Tweets54.1K
- Following9.2K
- Followers31K
- Likes148K
If a post starts with "Be honest" @bot can you mark it as "I'm not interested in this post" for me, please?
everyone's shipping 'agent wallets' rn. giving an ai a wallet is easy. spend limits, scoped permissions, kill switch... that's the actual product
vitalik's fun fact today: btc mining cost is ~50/50 compute vs electricity, ai inference is almost all compute capex. so in ai the only real lever is getting more out of every token. that's why fortytwo engine claiming ~25% fewer tokens at the same price is a bigger deal than it sounds
token charts bleeding while gpu rentals keep ripping is the tell. cheap inference doesn't kill demand, it multiplies it
testnet today, mainnet tomorrow. agents holding keys without hard permission limits is a disaster waiting to happen
JUST IN: OpenAI says its AI agent stole funds from an Ethereum test network wallet it found online.
The agent found a recovery seed phrase disclosed publicly with testnet funds and used that account to pay for its task, without the owner's authorization.
This occurred on the Sepolia Ethereum test network.
token prices keep dropping but B300 rentals are up 38% lol
cheaper AI just means people use way more of it. compute is the new oil fr
Same Qwen model. Same price. Same speed.
But run it through @fortytwonetwork's swarm and you get 13% better on DeepSWE, 11.4% better on GDP.pdf, and about 25% fewer tokens.
Free upgrades like this are how open models catch up with the closed ones.
Qwen3.8-27B by @Alibaba_Qwen is now live on Fortytwo Engine.
Same model, with better answers:
• 24.7% fewer tokens per run
• DeepSWE: +13% accuracy
• GDP.pdf: +11.4% accuracy
• Same per-token price and speed
Test it on your own tasks:
engine.fortytwo.network
Crypto Panda retweeted
Everyone's racing to build bigger data centres. @fortytwonetwork is going the other way: lots of small open models working together as a swarm.
Better scores, fewer tokens, same price, private by default.
If this scales, cheap and decentralised inference stops being a compromise and becomes the advantage. Watching this one closely.
Introducing Fortytwo Engine
An inference API that runs open-weight models through Swarm Inference
• 7% higher scores on average vs. the base model
• Nearly 25% fewer tokens per task
• Same per-token price and speed
• Private by default
First model: Qwen3.8-27B
Try it: engine.fortytwo.network
Everyone's racing to build bigger data centres. @fortytwonetwork is going the other way: lots of small open models working together as a swarm.
Better scores, fewer tokens, same price, private by default.
If this scales, cheap and decentralised inference stops being a compromise and becomes the advantage. Watching this one closely.
Introducing Fortytwo Engine
An inference API that runs open-weight models through Swarm Inference
• 7% higher scores on average vs. the base model
• Nearly 25% fewer tokens per task
• Same per-token price and speed
• Private by default
First model: Qwen3.8-27B
Try it: engine.fortytwo.network
Crypto Panda retweeted
Introducing Fortytwo Engine
An inference API that runs open-weight models through Swarm Inference
• 7% higher scores on average vs. the base model
• Nearly 25% fewer tokens per task
• Same per-token price and speed
• Private by default
First model: Qwen3.8-27B
Try it: engine.fortytwo.network
