J.D. Salbego

@fmerian Supabase is incredible, and not just because of the product's utility and features. What sets it apart is how accessible the entire experience is. The UI, the configuration, the documentation. Everything is designed so that even non-technical teams can navigate it without needing an engineer on call. We use Supabase as our database layer at ClawSecure and it's been one of the most...

@fmerian I'm a die-hard opus 4.6 extended thinker for everything. Why use a college grad's brain when you have Einstein's brain at your fingertips? But yes, it's incredibly expensive, but when you calculate the output value vs. the costs, it's a no-brainer.

@busmark_w_nika I would definitely go with the latter, based upon the scrutiny they faced for many years. As well, they still have a pretty negative perception within the global public and many areas and many different demographics of users. So this approach is a great way to spread information at a mass level to fix their public perception....IMO of course ;)

@busmark_w_nika The AT Protocol is genuinely interesting infrastructure regardless of what happens with Bluesky's user numbers. An open, decentralized social protocol that anyone can build on is the right long-term architecture. The problem is that most users don't care about protocol-level innovation. They care about where their friends are and whether the content is good. The X mirroring...

Cool concept and very honest about the conversion challenge with crypto payments. The MCP server integration so agents can search directly from Claude Desktop is a nice touch. Curious to see if the x402 payment model gets more traction as more agent-native commerce emerges.

@busmark_w_nika Hex Security stands out immediately. AI agents that continuously hack your system to find vulnerabilities is the offensive side of exactly what we're building on the defensive side. We secure the agent supply chain before and after deployment. They stress-test the systems those agents connect to. $1M+ run-rate in 8 weeks tells you everything about market demand. Security for AI...

@fmerian Claude Opus 4.6 with extended thinking for everything. Full stop. We run it across virtually every workflow, coding, security analysis, content, strategy, architecture, all of it. The only exceptions are basic boilerplate and simple admin tasks. The cost premium is worth it because the accuracy and output quality eliminates the rework cycles you'd burn through with cheaper models. You...

@rohanrecommends ROS setup is genuinely one of the most painful developer experiences out there. The config file maze, dependency conflicts, launch file debugging, simulator integration. It's the kind of problem where you spend 80% of your time on infrastructure and 20% on the actual robotics. If Drift can flip that ratio, the value proposition is immediately obvious to anyone who's fought with...

Having AWS-scale infrastructure behind browser agents changes the reliability equation. Parallel browser sessions with human oversight is exactly what production use cases need. Interested to see how this compares to the open-source alternatives in terms of developer experience.

LLM cost governance is something I think about constantly. When you're running multiple agents across different providers, spend can spiral fast without good routing and quotas. The x402 payment integration is an interesting twist. Are you seeing much adoption of the crypto payment rail from agent builders?

The 'private AI' angle is important. A lot of teams want agent automation but can't send their data to external services. Connecting to 120+ apps while keeping everything governed is a hard problem to solve well. How are you handling credential management across that many integrations?

Turning SOPs into actual executable automations is a really practical use case. Most teams have processes documented somewhere but the gap between 'here's how we do it' and 'here's an automated version' is usually weeks of engineering. The approval-step workflow is a nice safety net too.

Auth is one of those things nobody wants to build but everyone needs. Especially with agents that need to interact with 5+ services on behalf of users. The fact that you support 15+ frameworks out of the box makes adoption so much easier. Smart infrastructure play.

78K GitHub stars says it all. I've been exploring browser automation for some of my agent workflows and the approach of extracting interactive elements so the agent doesn't have to figure out the DOM is elegant. Simple API, powerful abstraction. The traction here is very well earned.

This solves a real pain point. The gap between 'agent can read my docs' and 'agent can safely query my production database' is massive, and most people just hack around it. Sandboxed views with row limits and query guardrails is the right approach. Really well thought out.

Been tracking Langfuse for a while and the growth is impressive. LLM cost tracking alone is worth the setup. I run multiple LLM providers across different workflows and having one dashboard to see token usage and latency across all of them would save me from some surprise bills. The self-hosting option is a big plus.

The visual observability piece is what caught my eye. I use n8n for some of my automation workflows and the ability to see what's happening at each step is invaluable. Bringing that same visibility to code-first agent frameworks is a smart move. How are you handling tracing for deeply nested agent chains?

This is exactly the infrastructure layer that's been missing. Everyone talks about building agents but nobody talks about 'what happens when you have 50 of them running and one goes rogue at 3am.' The zero-trust IAM for agents feels like it should be table stakes. Great positioning.

The self-healing automation code is a huge differentiator. I've dealt with browser scrapers that break every time a site updates their CSS, and having the AI fall back and fix the code automatically would save so much maintenance time. 15K+ stars is well deserved. Really impressive evolution from v1.

Those download numbers speak for themselves. The Studio UI for dev and testing is a great call. I've been building multi-agent orchestration systems and the hardest part is always observability. Being able to visually trace what your agents are doing during development changes the debugging experience completely.