News & Intel
LiveDaily AI-curated intelligence on Palantir Foundry, Ontology, AIP, Apollo, contracts, and community feedback. Updated automatically via GitHub Actions every day at 7 AM UTC.
Frozen Speedrun AIP Workflow Deployment
I just started learning to use Foundry and am getting stuck at the Deploy step of Transforming the Data. Every time I deploy, even after I have deleted and rebuild the entire pipeline from scratch, my deployment gets stuck. I have tried it both with the data outputs grouped and ungrouped by changing settings in the Use LLM node. Has anyone encountered this before or have an idea of a workaround? I can’t find anything similar on any of the tutorials or youtube videos. . 1 post - 1 participant Read full topic
IBM Jumps 5% as Bob Coding Agent Gains Self-Hosted Deployment; Salesforce Rises 3%, Palantir Inches Higher
The post IBM Jumps 5% as Bob Coding Agent Gains Self-Hosted Deployment; Salesforce Rises 3%, Palantir Inches Higher appeared first on 24/7 Wall St.. Enterprises that keep their source code behind their own firewalls now have a way to run Bob, the agentic soft…
NVIDIA Launches Open Agent Safety Platform to Secure Agents From Testing to Deployment
NVIDIA today announced NVIDIA Open Agent Safety Platform, an open software platform and reference system design to strengthen AI security from agent testing to deployment, with full-stack governance and control across software and the hardware, compute and ro…
Billions Behind the Autonomous Machine Boom as Unmanned Ground Vehicles Disrupts the Defense Industry
From battlefield logistics to mining and infrastructure, unmanned ground vehicles are moving rapidly toward real-world deployment NEW YORK, Sept. 24, 2026 /PRNewswire/ -- Platinum Trade Alerts News Commentary - Unmanned Ground Vehicles (UGVs) are quickly beco…
Harvey Puts a Former Practicing Lawyer in Every Deployment. About 180 of Them. Here’s How That Model Works
Everybody in B2B AI is rediscovering forward deployed engineers right now. Palantir ran the model for two decades, OpenAI built a team, and everyone has figured out agents don’t deploy themselves effectively. Not well at least, not yet at least. Harvey runs t…
Germany is evaluating 30 AI tools for its military with a prototype due in 2027
The Bundeswehr is assessing 30 artificial intelligence tools that could help speed up battlefield decision-making by processing data from drones and satellites. A prototype is expected in the second quarter of 2027, with full deployment planned for 2028. Vice…
The USS Abraham Lincoln Is a Troubling Metaphor
Like the Iran war itself, the carrier’s deployment lasted far longer than Trump expected or planned.
Installing Published OSDK in Pilot App
I’ve been using Pilot to prototype a front-end application against deployed ontology entities. The workflow so far has been: Defined object types, actions, and link types in the local ontology definition Pilot generated a local .osdk directory after deployment Built out the full UI importing from this local .osdk against production data However, if I want to move this towards a production deployment, I thought that generating a semantically versioned OSDK package via the Developer Console and installing it as an npm dependency would be better than keeping the OSDK local here. I’ve found (and I’m not sure if I’m doing something incorrectly) that if I then create the OSDK in Dev console and try to replace the .osdk with an import, I can’t easily give the pilot repo access to install the package. Is there a supported way to install a published SemVer OSDK package within a Pilot workspace? If not, what is the recommended path from a Pilot prototype to a production
Architecture question: is there a residual enforcement gap after Ontology permissions and actions?
Context: I am researching safety architecture patterns for AI/agent actions with consequential or hard-to-reverse external effects. This is a general design question, not tied to a specific deployment. I’d appreciate a reality check from people familiar with AIP and the Ontology. My understanding is that Palantir already provides strong software controls around identity, permissions, Action Types, submission criteria, and governed access. I’m not assuming there is a missing feature here. The narrower question is what happens after those controls have correctly authorized an action. Conceptually: agent/ application → identity & permission → Ontology Action / policy checks → authorized action → external system or consequential effect Is there any meaningful class of use cases where it is useful to separate “the software stack has authorized this action” from “this action is now allowed to actually take effect in the external system?” For example, would a separat
Superrepo Permissions Errors
i find i run into this problem often, I create the marketplace store, do a deployment, and despite being the owner, i run into these permission issues 1 post - 1 participant Read full topic
DC residents debate safety gains a year into Trump’s Guard deployment
Mixed reactions persist in Washington, DC, a year after Trump deployed the National Guard to address crime concerns.
Federation layer, Self-hosting of LLMs & Routing - More examples available?
Hello, I came across this interesting (but short) documentation regarding the implementation of a proxy or federation layer (I think mostly call model router outside the platform?). Bring your own model • Build a proxy or federation layer with AIP • Palantir While there is a tiny example mentioning the deployment of a container, additional documentation would be welcomed ? For example, we successfully deployed a 4B model asset as a container with vLLM engine but one thing we couldn’t do is to have the weights of the model loaded from a dataset ? Today, the weight are backed in the image. As no persistent storage is available, at every spin up we need to reload everything which takes around 5-10min while it would be better to have the model weights, CUDA graph etc.. stored in a dataset. This would allow: Lighter docker image Possibility of hot swap/sleep of model weights maybe GPU Release ? and more… Regarding the routing, where the router layer would be deployed
Palantir raises outlook as US commercial AI demand accelerates, sharpening sovereignty strategy
Palantir Technologies raised its full-year revenue and profitability forecasts after reporting a sharp acceleration in commercial demand, particularly in the US, underscoring the growing role of AI deployment and data sovereignty in enterprise software spendi…
Is it possible to rollback a schedule to an earlier version ?
*Hello Community Team, I’d like to know if there’s a way to revert to a previous version of a scheduler?,* because we have large schedules in our production environments whose configuration often breaks after deployments using templates. We got the case today and It would be great to have a simple way (via API or otherwise) to revert to a previous version of the scheduler or a specific version. 1 post - 1 participant Read full topic
Why are companies on the hunt for Forward Deployment Engineers? Explained
Forward Deployment Engineers are essentially software engineers who, instead of building software, work directly with clients to integrate and customise the software to their own needs
OpenAI and Datadog leaders back AI deployment startup Arrakis
London-based AI deployment startup Arrakis has emerged from stealth, having raised nearly $40m in just over three months. Arrakis was founded by former Accel investor Rafael Quintanilla, along with former Palantir executive Haroun Beltaifa, Romain Fouilland, …
When should you adopt the forward-deployed model?
Explore the forward-deployed engineer role, its origins, core responsibilities and why the forward-deployed engineering model is becoming essential for AI deployments.
OpenAI’s deployment arm plans to acquire Northslope as enterprise AI push accelerates
OpenAI's strategic acquisitions signal a robust expansion into enterprise AI, potentially reshaping industry standards and competitive dynamics. The post OpenAI’s deployment arm plans to acquire Northslope as enterprise AI push accelerates appeared first on C…