Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Introducing programs in Slate The first way to actually engineer your agent orchestration: Fable + hoards of GLMs Deep Research + Kimi 2.7 Auto-research any task Design reusable components that combine to build your system Even bring your own subscriptions

83,367 Aufrufe • vor 2 Monaten •via X (Twitter)

24 Kommentare

Profilbild von fAi S.
fAi S.vor 2 Monaten

Omg this is amazing! 🔥

Profilbild von akira
akiravor 2 Monaten

Thank you!!!

Profilbild von Oliver Beavers
Oliver Beaversvor 2 Monaten

Hell. Yes.

Profilbild von akira
akiravor 2 Monaten

Trust me. It rips. And because it works with any model You can actually just rip it with Kimi and GLM

Profilbild von Oliver Beavers
Oliver Beaversvor 2 Monaten

Love it. Excited to see more

Profilbild von AI Mastery Guide
AI Mastery Guidevor 2 Monaten

Reusable components for agent orchestration is the direction this space needed to go 🔧

Profilbild von akira
akiravor 2 Monaten

Been needing to for a long time

Profilbild von AI Mastery Guide
AI Mastery Guidevor 2 Monaten

Feels like the whole space is finally catching up to this idea

Profilbild von akira
akiravor 2 Monaten

I feel like you would like using it, try it out, lmk how it goes $5 credits and you can use your codex sub npm i -g @randomlabs/slate

Profilbild von AI Mastery Guide
AI Mastery Guidevor 2 Monaten

Definitely trying this out, curious how it feels in practice 🙌

Profilbild von daksh
dakshvor 2 Monaten

oh hey I once met you at a potluck

Profilbild von akira
akiravor 2 Monaten

Try it out: npm i -g @randomlabs/slate

Profilbild von thud
thudvor 2 Monaten

Released today?

Profilbild von akira
akiravor 2 Monaten

Yes sir A glorious day indeed

Profilbild von sof.
sof.vor 2 Monaten

does the deep research component work with any glms or just kimi

Profilbild von akira
akiravor 2 Monaten

Yeah works with any model You can run deep research with 5.2, deepseek, etc

Profilbild von kiosa
kiosavor 2 Monaten

fable plus hordes of GLMs sounds illegal but useful

Profilbild von akira
akiravor 2 Monaten

It’s insane actually my velocity is through the roof npm i -g @randomlabs/slate

Profilbild von Peter Petrash
Peter Petrashvor 2 Monaten

v cool. open source?

Profilbild von akira
akiravor 2 Monaten

Not open source but works with your subscriptions and BYOK

Profilbild von Samian
Samianvor 2 Monaten

calling it "engineering" when it's just picking models and gluing them together is a stretch. orchestration starts where that UI ends

Profilbild von akira
akiravor 2 Monaten

Wrong I would have previously agreed, running a REPL for orchestration is not engineering A programmable runtime, with guarantees, namespaced memory, and an execution model is more like a VM With that, it allows you to write software and engineer actual systems Id argue engineering is mostly data modeling, system decomposition, and runtime tooling. This has the ability do all of it. The first actual way to author and engineer your system like normal software

Profilbild von Samian
Samianvor 2 Monaten

fair point — a proper runtime with memory namespacing and guarantees is more legit than I gave credit for. what stack are you building on?

Profilbild von akira
akiravor 2 Monaten

TS and everything basically built in house

Ähnliche Videos