@kunchenguid Muse uses fastly/cloudflare relay, while instinct uses raw GCP.
I expect most large-scale use cases will end up adopting Museโs approach. No serious cloud provider will put up with their IP being tainted like this if it happens at scale.
@borrowed_ideas This is closer to free model given its price
You can look at glm 5.3 flash (oxAlpha) share when it was free vs now - there is a very large difference.
Will this stay where it is when its not discounted
@Merridew__ I don't get finance folks - do they ever use these models. All these 2nd tier models - grok, muse spark, chinese models look good on questions because they can compare it to whats actually good - fable/gpt/maybe gemini. But they'd never actually use them as their daily driver.
@MetacriticCap@evrgn11112231 I donโt think this distinction is that helpful.
Claude Code/Codex are themselves agentic workloads. And at least in subagent/exploration flows, these systems already use or support cheaper models for narrower tasks (like haiku in claude code)
@Alice_comfy This can be adapted to intraday shorting too.
From backtesting a simpler strategy like short once and close EOD is better. However, rebalancing is more execution friendly - since you can maximize on the few shorts you managed to get in (for the locate poor).
Remote Agent is live in VS Code (v0.472.1).
Early users are already shipping while they sleep. One dev launched 3 agents before logging off - woke up to a PR, docs, and a fixed flaky test.
Weโre celebrating wins like that by giving 5 devs 4 months of Augment Max ($1K in value each).
RT + reply with your worst backlog task. Every 100 RTs = one more winner
Remote Agent is live in VS Code (v0.472.1).
Early users are already shipping while they sleep. One dev launched 3 agents before logging off - woke up to a PR, docs, and a fixed flaky test.
Weโre celebrating wins like that by giving 5 devs 4 months of Augment Max ($1K in value each).
RT + reply with your worst backlog task. Every 100 RTs = one more winner
@KrisVerma88@OceanSecMktg How do they compare against center point ? Center point just feels like a more modern system overall - with a proper API support. With Cobra/Ocean you have to use DAS which feels clunky.
@Makeabuck2 I think there are aspects of it that are more comparable though - amazon can be the first place a person goes to search for something (a bit like google), or ad placement in prime video.
@Makeabuck2 In competitive categories where fixed % fees are limited by low prices, auction-based ads let them extract higher fees since sellers bid aggressively for visibility. It's not just another ad platform like Google/Meta - it's a more efficient marketplace fee model.
@MetacriticCap Thank you for saying this. Everyone on my timeline thinks nothing changes - but i feel like it has to given new information and a nuanced take is better than just denying everything.