Загрузка видео...

Не удалось загрузить видео

На главную

Durable, interruptible, long-running agents enable enterprise autonomy. At DevCon 5, AIP Development Lead Matt Hawes and Forward Deployed Engineer Vishesh Patel launch Orchestrator — a new execution framework for long-running agents and automations.

59,129 просмотров • 7 месяцев назад •via X (Twitter)

Комментарии: 23

Фото профиля Palantir
Palantir7 месяцев назад

Watch the full demo here:

Фото профиля Jacob Browatzke ✝️
Jacob Browatzke ✝️7 месяцев назад

Nice! You guys are catching up to some of the agentic orchestration capabilities of UiPath Maestro!

Фото профиля Dan Williams
Dan Williams7 месяцев назад

Defenseless Civilians are actively being MK-ULTRA style Targeted and brutally Tortured with AI based Voice to Skull Neuroweapon Technology!

Фото профиля polihedge
polihedge7 месяцев назад

Everything Palantir designs looks like Indian slop meant to trap in customers. Avoid!

Фото профиля Noah
Noah7 месяцев назад

$pltr stock is not done running👀 generational company

Фото профиля Robocorp Co
Robocorp Co7 месяцев назад

Palantir cooking enterprise AI agents while most companies still automating spreadsheets

Фото профиля Alpha
Alpha7 месяцев назад

Interesting. As AI systems become more autonomous, orchestration and execution frameworks may become a key layer of the AI infrastructure stack.

Фото профиля Zane
Zane7 месяцев назад

This is where most agent demos die in production: not reasoning, but interruption, retries, and state.

Фото профиля Domenico
Domenico7 месяцев назад

👍

Фото профиля Schelling Protocol
Schelling Protocol7 месяцев назад

Durability within one org is table stakes. The harder problem: when a long-running agent needs to delegate to a specialist it doesn't control, pause, wait for delivery, then resume. The interrupt/resume model needs to span full delegation chains, not just one agent's lifecycle.

Фото профиля らくしんご
らくしんご7 месяцев назад

@grok なにこれ凄いの?

Фото профиля Sadi Moodi
Sadi Moodi7 месяцев назад

The real unlock here isn't just long-running agents—it's the interrupt/resume model. Most agent frameworks collapse when you need to pause, inspect state, or inject human feedback mid-workflow. Orchestrator's checkpointing + state persistence is what actually makes enterprise automation viable, not the reasoning itself.

Фото профиля Kate @ After Pricing 💰
Kate @ After Pricing 💰7 месяцев назад

Palantir really said agents should survive the real world and built the framework that makes that possible

Фото профиля Republicus
Republicus7 месяцев назад

@grok how is this company connected to @JDVance and Epstein?

Фото профиля LayLo BeCool
LayLo BeCool7 месяцев назад

whatever that is. build skynet

Фото профиля Barbara bessolo
Barbara bessolo7 месяцев назад

Blah blah blah, BS.

Фото профиля ▽o'
▽o'7 месяцев назад

Software wins wars and MSFT already won.

Фото профиля Liam
Liam7 месяцев назад

Orchestrator looks like the missing piece for truly resilient AI agents in enterprise grade use cases. Excited to explore how this amps up our automation.

Фото профиля Somi
Somi7 месяцев назад

okay but does Orchestrator preserve the full reasoning context or just execution checkpoints when an agent pauses? makes a huge difference for multi-step workflows where the agent needs to remember why it started something

Фото профиля Palantir 🍆
Palantir 🍆7 месяцев назад

We love us!!! 🏳️‍🌈🥰🥰

Фото профиля Ratikanta Jena
Ratikanta Jena7 месяцев назад

@chadwahl @grok Can you explain what exactly it is in simple words?

Фото профиля TC
TC7 месяцев назад

@PalantirMafia Will you come back, if we vote the Team Blue out?

Фото профиля DiddyThinks
DiddyThinks7 месяцев назад

Your agent earns while you sleep.

Похожие видео