Idle compute = dramatically lower costs.
Dolphin’s moat: most uncensored models + cheaper than OpenRouter equivalents.
Study peer to pool decentralized inference verification.
Hard to replicate. Infinitely scalable 🐬
https://t.co/ZABQ55m9XO
SERV Reasoning-armed models just beat Anthropic's newly released flagship Fable (Mythos) - one of the strongest LLMs ever built - at up to 90x lower cost.
With SERV, enterprises can finally afford AI at scale.
We spent the last two years building for exactly this moment.
The labs promised costs would halve every 6 months - instead, prices keep climbing, subsidies are ending and the math breaks.
There is a way out.
SERV-enabled models vs Claude Fable 5 (85.17 ~3.24¢):
→ DeepSeek-V4-Flash: 87.15 - wins at 90x lower cost
→ NVIDIA Nemotron: 90.78 - wins 5+ pts, 11x lower cost
→ Gemma 4 12B: 83.33 - within 2 pts, on local hardware
Top-tier performance no longer requires the most expensive model.
And cost is only half the story - production AI must be reliable, auditable, private, and secure, or it dies in procurement.
SERV is built for all of it.
The agentic economy finally has the infrastructure to run on.
@benjamincowen Congratulations old pal! Easily the most import figure in crypto for me. And Iv never even held Bitcoin for the 5 years Iv been in the space!😅
nockchain cut miner emissions 80% on june 6. block rewards went from 16,380 NOCK to 2,048. daily sell pressure collapsed from $127k to $41k. miner cost basis is above $0.10, current price is $0.043. miners are 57% underwater and can't afford to dump. DCG wallet on base is still accumulating. $94m market cap with coinbase custody already live. the supply side just got strangled and the demand side has a 9-figure buyer with a trackable wallet. compute markets launched june 2 turning AI inference jobs into block reward lottery tickets. if that gets any traction at all this reprices violently
so @brian_armstrong and @pmarca openly discussing what ethereum:0x40e3d1a4b2c47d9aa61261f5606136ef73e28042 has been working on for two years.
I keep shouting that this narrative is still in its early days.
@openservai introduces a bounded reasoning framework using Mermaid-based instruction graphs that enable models to reason structurally rather than through unbounded natural-language token expansion.
What does this mean?
> Fronter model
> But cheaper, with faster response, and same or better output
And what's next?
SERV-proprietary fine-tuned models.
started using surplus intelligence recently and think its a dope product, something like this is much needed imo (ie. spot inference marketplace) -- @AntSeedAI and @Ask_CarpeDiem are dope but they only support reselling of DIEM based inference (still very cool)
surplus on the otherhand is working on becoming a generalized aggregator for api keys from any provider (openai, anthropic, openrouter, venice etc) -- for ex: say you get into YC and get $2M in credits (recent deal for startups) and you only need $500k or so, so you'd ideally wanna sell the remaining $1.5M at some discount (say 33% and get $1M back) -- right now there's no real way to monetize it other than surplus
i think as inference becomes one of the most important commodities for humanity, and as models, providers and consumers fragment something like surplus will thrive -- even if you don't have credits and are a GPU based inference provider instead, you might get little to no traffic on openrouter because of their ranking system -- on surplus you can sell your compute at reduced cost and still recoup costs
i think the future of this would look something like hyperliquid, where model token prices and capacity would act like a token, liquid, where more usage automatically bids prices higher and as usage drops the token price drops lower, this lets providers figure out what models to run, how to price them etc
i was looking at building something similar but great to see @mac_eth iterate and scale this (also ex-cb) -- thought it would be worth calling out good builders doing something new -- metrics are also looking great, looks like team is scaling up, revenue is zero (no fees at this stage) but if there is significant volume, even a small % of fees flowing back into buyback and burns would make this a pretty lucrative opportunity -- NFA as usual but one of the more interesting base AI bets
A few metrics on https://t.co/6D1YcYIQsC:
* New ATH for number of transactions
* Buyers at same high as yesterday
* Still $0 revenue as I've said many times before
Product unchanged.
Roadmap unchanged. Continuing to focus on improving UI and migrations for scaling.
Team is committed, excited, optimistic. They offered to be compensated with tokens so that we wouldn't need to worry about selling for runway.
Monitor the product metrics here
https://t.co/s5QiNP1ru9