emo.eth

exploring agents. prev: stablecoins at @nativemarkets/@stablecoin/@stripe, consumer/social crypto, @opensea, seaport co-author | he/him | views are your mom's

[emo.eth @emo_eth]

Um, actually, Ethereum is the name of the protocol, not the cryptocurrency

You're thinking of "Ethereum's Monster"

[emo.eth @emo_eth]

@0x_Osprey @secjack_ ok i’ll google what serverless and edge mean… it’s time. i’m ready

[emo.eth @emo_eth]

i don't know what "durable objects" are and at this point i'm too afraid to ask

[emo.eth @emo_eth]

@secjack_ @0x_Osprey i’m like three tech eras of jargon removed, i never learned what most of these terms mean 😔

[emo.eth @emo_eth]

@NadavAHollander i’m still working on regular object permanence tbh

[emo.eth @emo_eth]

edge functions also one of those buzzwords i never learned but iiuc you’re saying durable object is like every object has its own personal datastore that it can only write to serially
as opposed to parallel writes to central / non-collocated db/storage service
so like a tardigrade or pi durable sqlite session - and that’s why good for agents?
and fast to start when needed
ok cool i think i get it, ty

[emo.eth @emo_eth]

@arjunkmrm none of these words are in the bible

[emo.eth @emo_eth]

@johnpalmer congrats! i know the team was very excited about link before i left

[emo.eth @emo_eth]

got an M5 Ultra 256GB Mac Studio that i need to decide if i want to keep

hit me with benchmarks you wanna see

[emo.eth @emo_eth]

“stop doing something stupid” still undefeated for large-multiple speedups

[Dylan Garcia @_dylanga]

protip: make sure your database indices are being used

[emo.eth @emo_eth]

@mikeroySoft @0xSero @ornith_ ty, was looking for this

[emo.eth @emo_eth]

there's no reason to run glm-5.3-flash when deepseek-4.1-flash-1017 is both smaller and smarter, right?

[emo.eth @emo_eth]

i do think there is a time and place for skills, though

i don't think they should ever be auto-loaded

[emo.eth @emo_eth]

nothing has wasted more of my agents’ (and my) time more than unnecessary skills

[Tak 🦞 @cherry_mx_reds]

Anthropic is now recommending that companies REMOVE unnecessary subagents from their AI systems.

This is a pretty significant shift in agent architecture, with their team finding that a single agent equipped with skills can frequently outperform complex multi-agent systems, and one customer seeing roughly 2x faster execution.

The emerging rules of thumb:

→ Skills: Teach the agent HOW to do something, including procedures, expertise, and instructions.

→ Tools / MCP: Give the agent the ability to DO things through APIs, actions, and live data.

→ Subagents: Use for parallel work, isolated context, or tasks a single agent struggles with.

The interesting part is progressive disclosure. Instead of spawning specialized agents or stuffing everything into a giant prompt, one agent discovers and loads skills and tools as needed.

Their advice is to fold existing subagents into the main loop, turn their instructions into skills, expose their actions as tools, and run evals to see if performance holds up.

As models have gotten smarter, improving agent architecture is becoming less about what you can add and more about what you can delete.

https://t.co/DNq60Skkmc

[emo.eth @emo_eth]

@backnotprop agreed, trying to figure out something that works for me. would love to chat about this sometime!

[emo.eth @emo_eth]

TIL “leech” and “leech” are not only spelled differently but have different etymologies

huh. guess i just figured that leach was a metaphor for, like, sucking minerals or chemicals out of something

[emo.eth @emo_eth]

this bout to be me

[emo.eth @emo_eth]

restricted my coordinator agent's tools to just ask, task, and wait to force it to delegate all work, since it steadfastly refused to listen to instructions to delegate everything

it just started using the cua-driver mcp to do stuff instead 😐

[emo.eth @emo_eth]

@binji_x what besides agentic payments/x402 is "exciting" rn? agree the substrate is there but i don't see anything obvious to build that stripe won't eat

[emo.eth @emo_eth]

been meaning to do this

my goal is to have all my tailnet devices (+ ideally remote provider apps like grok or chatgpt) sync all sessions to a single devices, which then indexes them and makes them searchable via qmd or zvec

can both act as memory and training data then

[Ben Holmes @BHolmesDev]

We should stop keeping all of our agent conversations local. They should be stored in a shared place.

This lets you do some powerful things:
- Schedule self-improvement agents to review runs and pull out failure modes worth fixing
- Replay past conversations against different models for benchmarking

I met a team of devs zipping up Claude and Codex sessions from their computers and shipping them to a shared S3 bucket. That's... a start 😄 But really, all agent sessions should happen on a single source (like a cloud runner) so that conversation traces are stored together by default.

It's why Warp Factories uses cloud sandboxes and makes sessions available via API. We also added an inbox view to see which conversations need your attention, which is quite convenient:

[emo.eth @emo_eth]

idk why but i like that english has a lot of phonemes. a lot of its vowels don't exist in other languages. that's cool

its spelling could use an overhaul though