"The world doesn't need AGI, what it needs is domain-specific GPT-4 level models." Check out this interview with our CEO @TylerXsaltsman on why the world doesn't need AGI, smaller domain-specific models will dominate, and how the future won't be one big mega-model. finance.yahoo.com/technology…

We're hiring to support our growing work with the U.S. military and national security partners. Research Scientist – you will join the EdgeRunner Research organization to train models, evaluate AI systems and harnesses, and bring world-class AI to the warfighter. A PhD in a relevant field is preferred, but we will also consider relevant research experience from time in industry, military, or government. ML Engineer – you will build the foundations of the EdgeRunner Research organization. Behind every successful model is code for data acquisition, data processing, model evaluation, efficient parallelization, high-performance kernels, low-loss quantization and compression, and much more. Candidates must be eligible to receive a TS/SCI clearance, and strong candidates will have an active clearance. Under federal law, U.S. citizenship is required to obtain a security clearance. Learn more and apply here: jobs.ashbyhq.com/edgerunner-…

Great to be on Capitol Hill this week meeting with lawmakers to discuss the importance of Open Weight models for American AI Leadership. We also celebrated the launch of the bipartisan House Innovators Caucus to support the Little Tech agenda. Thank you @Innovators and @a16z for organizing.

We're heading to the @AFA_Air_Space Association Conference! We'll be showcasing how our AFSC-specific AI agents support military operations in contested environments. Dates: September 14-16 Location: Gaylord National Resort & Convention Center Booth: MB313 Learn more and register here: afa.org/air-space-cyber-conf…

We're excited to share additional results on repetition loop reduction in our latest research paper, "When Can One Neuron Fix Repetition Loops in LLMs?" In this update to our previous work, we (1) compare our method against Liquid AI's recently shared method for solving the same problem and (2) test our method not only with Gemma 4 models but also the LFM and Qwen model families. Read the paper below or access it on @arxiv here: arxiv.org/abs/2606.13705

We're heading to NGAUS 2026! Alongside our partner @HP, we'll be attending 148th General Conference & Exhibition the hosted by The National Guard Association of the United States (@NGAUS1878). Dates: August 28-31 Booth: 1732 (@HP) Location: Indianapolis Convention Center Learn more and register here: ngaus.org/events/148th-gener…

Our CEO Tyler Saltsman joined @MorningsMaria on @FoxBusiness to discuss EdgeRunner-Camo, our new domain-specific LLM for the warfighter developed in collaboration with the US Army Artificial Intelligence Integration Center (AI2C). Watch here: vimeo.com/1219433753

Join us now @MorningsMaria I'm in for @MariaBartiromo and we've got a big show ahead w/@TIFFANYMSMILEY @SAMIRNKAPADIA @CHERYLCASONE @REPRYANZINKE @LEONCFRA @DANIELHOFFMANDC @REPMIKELAWLER @SONENSHINEJ @TROYGAYESKI @EDGERUNNERAI @BEARTRAPSREPORT @BNYglobal @MadisonAlworth

We're excited to announce EdgeRunner-Camo, a new domain-specific LLM for the warfighter from EdgeRunner AI developed in collaboration with @USArmyAI (AI2C). EdgeRunner-Camo achieves a 37% reduction in error rates across personnel, logistics, training, policy, weapons, and tactics, demonstrating how domain-specific LLMs can improve operational effectiveness. Domain specialization, including for military functional areas, allows us to shrink model sizes by 10x and more while still achieving frontier-level accuracy on military tasks. Such efficiency is important not only for enabling edge deployment on each warfighter’s local hardware, but also for radically decreasing costs when hosting AI solutions in server-client architectures. The work presented in this blog post represents our initial efforts with this dataset and use case. EdgeRunner-Camo is well suited as a generalist virtual assistant for the Army that can be deployed in-garrison on a wide variety of inference architectures or in the field at the tactical edge using EdgeRunner’s optimized inference runners. Read the full details below ⬇️

"We think of robots as the next frontier of AI" - Tyler Saltsman Check out our latest coverage by @FoxNews of our ongoing collaboration with @USArmy DEVCOM Army Research Laboratory to build and deploy localized LLMs for drones and robotics. Read the article here: foxnews.com/politics/battlef…

We're heading to @DoDIIS_WW! The theme of this year’s conference is DIA Next: Intelligence Technologies for Battlespace Lethality. Join us alongside our partner @HP to see how we're providing decision advantage with military-specific AI at the tactical edge. Dates: August 9-12 Location: Tampa Convention Center Booth: #1721 (HP, Inc.) Learn more and register here: ncsi.com/event/dodiis/

We're #6 for Mission Software & AI on The All-World Defense 300. The All-World Defense 300 by @_futureunion_ is the first authoritative review of the global private companies the allied world now depends on for its defense — ranked on outcome-based capabilities and demonstrated success in solving the challenges of customer nation-states. See the full list here: futureunion.co/top-defense-c…

Since launching last week, more than 230 companies and organizations from across the tech sector have signed the "Open Weights and American AI Leadership" open letter. We want to thank these partners for standing up and publicly supporting broader access to AI innovation. A special thanks to @nvidia, @a16z, and @PalantirTech for working with @Microsoft on this effort. These signatories understand that America’s AI leadership will not depend on the success of our frontier models alone, but on our ability to build a strong, secure, and open ecosystem that diffuses AI into every sector. We look forward to continuing to work with our partners and with policymakers to build that open ecosystem in a way that benefits American businesses, empowers American workers, and strengthens the American economy.

Today we are in the New Mexico desert alongside @DIU_x and @CENTCOM demonstrating how EdgeRunner enables real-time UAS/CUAS decision advantage and tactical intelligence for warfighters in DDIL environments. More to come - stay tuned.

By 2027, a destroyer XO should demand AI that operates fully offline without degradation, exposes transparent and traceable reasoning, integrates without destabilizing shipboard networks, reduces administrative and cognitive burden measurably, and fails gracefully, loudly enough that watch standers notice before the gap becomes a hazard. Check out this article from EdgeRunner’s John Babick in @CIMSEC: cimsec.org/what-a-destroyer-…