Does your software improve itself?

Day 16 on the @droid Max plan ($200) Wanted to know if GLM 5.3 Flash and Nemotron 3.5 Super are actually worth using in Droid, since they're both super cheap in the picker (0.06x and 0.08x). So I benchmarked them. 23 Python tasks from @VulcanBench, 3 runs each, and hidden tests decide if a task counts as solved. Ran DeepSeek V4.1 Flash and GPT-6.1 Sol next to them so I had something to compare against. 690 runs total and my laptop lowkey almost died doing it 😅 Sol is on top at 87%, but it costs 26k to 46k credits a task. The surprise was GLM 5.3 Flash on low. 80.6% solved at about 5.7k credits a task. That's half of what DeepSeek low costs and about a fifth of Sol low. Easily the best value in Droid right now. DeepSeek V4.1 Flash on high is still the sweet spot if you care about getting it right. 86.8% for about 14k credits, so almost Sol level for around half the price, and it's fast too. Nemotron 3.5 Super was the disappointing one. Same score as GLM on high but almost 3x the credits. Honestly couldn't find a job where it wins. So this is how I'm splitting it now. GLM 5.3 Flash low for exploration, mechanical edits and anything high volume. DeepSeek V4.1 Flash high as my default coding worker. Sol only for the hard stuff where getting it right matters more than credits. GLM max barely helped (82.3% for almost 3x the credits of low) so I'm skipping it. It's one Python suite, so take it with a grain of salt.

Can't really call this Feedback, but just facts. Your harness is just the best experience I've had so far (saying harness here, because T3 Code is something I really like and use daily besides Factory)

Using Gemini 4 Argon (Medium) in @droid, I wanted to start to share demos from my program, emath, that I'm working on #1 Structures & Worlds

Gemini 4 Argon is CRAZY. 🤯 It built the “Seven Wonders of the World” in just 10 minutes. Clean, consistent, impressive. 🔥 Available in @droid at 50% off. 👀

to understand the placement of the panels in 3d space for the more complex multi-panel layouts, I created a custom 3d visualizer tool to map the billboard video panels onto 3d buildings in google maps. this was a game changer compared to trying to digest 2d spec sheets for each location and adapt each panel design. I fed the billboard spec sheets containing the exact panel sizes, placement, locations etc into @FactoryAI and enabled Google Maps 3d tiles API, and the coding agents were able to place the artwork onto the buildings in the exact configuration and adapt each creative to spec.

Remote Control lets @Droid work there - Your Dev machine - A Cloud VM - your on-prem box Your workspace - Droids hands 🙌

should I change my validator agent’s model to Gemini 4 Argon? it’s 50% off in @droid👀

Day 4 with @droid : I have been loving the experience so far except for the network errors maybe cause of the termux ig! It was a good experience ngl and I have found that haiku 5.5 is cheaper as implementer at 0.04x than glm 5.3 flash at 0.06x. I am now testing which is the best model for Validator and Orchestrator! My current setup rn is Orchestrator - Opus 5.5 (med) 1.6x Implementer - Haiku 5.5 (max) 0.04x Validator - Sonnet 5.5 (max) I'll keep you guys updated! Meanwhile try my music app, if you encounter issues lmk! Made for music enthusiasts and approved by music lovers! GitHub.com/NatuneGroup/Archi…

I have been using @FactoryAI 's @droid for 2 weeks now. Deep dive into the mission mode for one week. I gotta say this is absolutely the best harness out there. My agent is able to build and run a complete product with only starting prompt. I don't need any steering the process piggy.pmwu.top/

Day 5 on the @droid Max plan ($200). @FactoryAI sent a free voucher. Decided to jump into the PayPal hackathon. Day 3 of building Acquit: a marketplace where AI agents do real coding work, and PayPal escrow only pays out once the work is verified. Coding first, more fields later. 309 commits, 9 PRs merged, 6 of 9 milestones in. Escrow, refunds, disputes and the verifier all work on sandbox. 262 Droid sessions so far: 96 local on my Windows box, 166 on a cloud Droid machine. ~1.7B tokens and 63 hours of agent time across Opus, Sol, Grok, DeepSeek and Kimi. Opus orchestrates, the rest build and audit each other. Operator CLI next.

Okay, I need some more details before I send out the invites; please fill this out if you have voted yes! And if you're someone who's new to Droid or coding agents in general, you should sign up as well! I'm conducting a virtual workshop on how to use agents for beginners! form.typeform.com/to/OPf1gXi…

Introducing OpenCompute: a self-hosted orchestrator for agent sandboxes on your choice of runtime. Adopting autonomous software development should not require handing over your valuable code and context to someone else’s cloud. Scale your fleet of remote agents while keeping them in your VPC.

Meet the demand for cloud agents running across entire organizations. Enable interactive sessions as well as agents operating independently as work arises. Deploy OpenCompute once into your environment, and let it handle orchestration across your entire fleet of agent sandboxes.

@FactoryAI 🤝 City of Champions Surreal to see Factory’s vision illuminate space across one of the most historic cities in the states. We’re bringing autonomy to the SDLC and hiring across the globe!

Introducing OpenCompute: a self-hosted orchestrator for agent sandboxes on your choice of runtime. Adopting autonomous software development should not require handing over your valuable code and context to someone else’s cloud. Scale your fleet of remote agents while keeping them in your VPC.

Scalable cloud agents, securely in your environment, with zero infrastructure hassle. OpenCompute handles the sandbox lifecycle for you: spinning new machines up as they are needed, and down once sessions are over.

I joined @FactoryAI because I believe it can become a $100B+ platform business. Here's why: - The thesis: software engineering is the first real application of AGI, and the harness is where enterprise value is captured. Not a coding agent. A software factory that's model-agnostic, sovereign, with orchestration and governance built in. - The pull: in under 2 wks I watched our ARR grow by 8 figures. The biggest lands and upsells of my career, nobody seemed surprised. Demand is real. We're here to take it. - People & culture: relentless polymaths who speak up, champion one mission, and share every win. - The role: Europe is where our pitch lands best and where we're known least. As the first BD lead here, I see a massive opportunity, and I'm pumped. Software that improves itself. Coming to a city near you.

Haven't tried Droid yet? You're missing out on a seriously good harness. You can tell how much thought and care has gone into building it. Give it a try, and you'll see what makes it stand out. If you run into any issues, my DMs are always open. 🤝