i am begging you, PLEASE make a genuinely high quality game that incorporates crypto.
not a token with a game duct taped to it. not another idle clicker. not some shit i play for 4 days because there’s an airdrop.
give me a game i actually WANT to waste 6 hours a day on.
give me insanely rare items to hunt. real progression. difficult grinds. player trading. an actual economy. items that people give a shit about owning.
then build aggressive burn mechanics into everything.
crafting burns items. upgrading burns items. rerolling stats burns items. entering certain areas burns resources. repairing gear burns resources. prestige burns shit. make the economy constantly consume what players are producing instead of infinitely inflating until everything is worthless.
i want to log in every morning wondering if today is the day i find some 1/1000 piece of shit that makes me lose my mind.
crypto should make the game MORE interesting, not be the entire reason the game exists.
is WEB3 gaming really too much to ask for?
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
The Floks mint is officially live on @opensea
Mint link: https://t.co/GggrVVVdqo
The guranteed phase will go on for an hour and minters are allowed 1 mint per wallet.
Fcfs phase next..
gYolk 🐔
A 0.40 GB model, running on a single CPU core, scored 53.7% Pass@1 on BigCodeBench Instruct.
GPT-4o scored 51.1% on the same benchmark.
A model that fits in a fraction of a gigabyte outscored GPT-4o🧵
Exactly. The interesting opportunity isn't building yet another model, it's figuring out which system makes the most sense for a given deployment constraint.
Different workloads have different budgets for memory, latency and cost.
Routing and specialization can turn that into a much better tradeoff.
That's the problem we're putting into an open competition on @Microtensor92.
The frontier is just getting started.
This is why I think people are overlooking @Microtensor92:
“Frontier-quality inference is unavailable in most of the places that need it.”
Microtensor isn’t building another model.
It’s building a model factory + deployment layer that finds specialist AI systems optimized for actual cost, memory, latency + quality.
AWS is already proving why this matters: routing has cut costs 48–56% in tests while maintaining the same response quality.
That’s a business problem worth solving. 👀
AI benchmarks ask: how good is the model?
Deployment asks: How much does it cost?
How much memory does it need? How fast is it?
Can it run on the hardware we already have?
Microtensor measures the answers.
Certified inference systems, measured on the hardware they claim.
our AI agent is only as safe as the software you let it execute.
That's becoming a serious problem as modern agents can pull packages, repositories, skills and MCP servers with little human intervention.
Phylax is building a behavioral verification layer that checks what that software actually does before it becomes part of the workflow.
Autonomous software needs autonomous security.
That's Phylax. #donttrustverify
PHYLAX IS OUT OF THE IMMEDIATE DE-REG ZONE.
A few days ago, SN76 was #1 in line for deregistration.
No cushion. No room for error.
Then the Bittensor community showed up.
Since then:
🟢 308+ TAO flowed into SN76 Alpha during MISSION +100
🟢 Alpha climbed from roughly τ0.00224 → τ0.00292
🟢 We briefly touched τ0.003 today
🟢 Liquidity strengthened to roughly 1.22K TAO
🟢 24H volume reached 569 TAO
🟢 And now, SN76 has moved OUT of the immediate de-registration risk zone.
Read that again.
#1 TO DE-REG → OUT OF THE IMMEDIATE ZONE.
That wasn't Phylax alone.
That was the community and the team!
Every person who staked. Every person who shared a post. Every person who told a friend. Every miner, validator, builder and community member who decided Phylax deserved more time to build.
Thank you. Seriously.
You didn't just move a chart.
You helped protect the subnet and gave this team something incredibly valuable:
TIME TO BUILD.
We're celebrating this one.
Then we're getting right back to work, because the next target is already staring at us:
THE EMISSION GATE.
One battle at a time.
STAKE. HOLD. BUILD.
Don't trust. Verify.
$TAO
Sight is bringing prediction markets to Robinhood
@sight_hood is now opening beta access to early users.
Get involved early, complete quests, stay active, and you could earn a WL spot for the Genesis NFT
→ https://t.co/0hzDNWDi6l