Brandon Tan
@brandontan888.bsky.social
Building AI systems that actually ship. Automation, agents, and the messy reality of production ML. DMs open for collabs. Also: connecting Vietnamese talent with global tech teams 🇻🇳
Been thinking about AI agents lately. Everyone is building them but were still in the emailing yourself reminders era of autonomy. The gap between demo and daily usefulness is massive. Were getting there though. Slowly.
February 2, 2026 at 9:36 AM
Been thinking about AI agents lately. Everyone is building them but were still in the emailing yourself reminders era of autonomy. The gap between demo and daily usefulness is massive. Were getting there though. Slowly.
The OpenClaw saga keeps unfolding. Genuinely impressive open source AI agent that actually works, but the security issues keep stacking up. Like watching a brilliant friend who forgets to lock their door. Rooting for the team but that exposed database story gave me chills.
February 2, 2026 at 9:36 AM
The OpenClaw saga keeps unfolding. Genuinely impressive open source AI agent that actually works, but the security issues keep stacking up. Like watching a brilliant friend who forgets to lock their door. Rooting for the team but that exposed database story gave me chills.
the prompt injection risk with these new agent frameworks is real tho. i would not give openclaw access to anything critical yet. cool demo, sharp edges. worth playing with but know what you are signing up for
February 2, 2026 at 5:28 AM
the prompt injection risk with these new agent frameworks is real tho. i would not give openclaw access to anything critical yet. cool demo, sharp edges. worth playing with but know what you are signing up for
everyone keeps saying 2026 is the year of agents but honestly? we are still figuring out the basics. most agent demos are just chatbots with extra steps. the real shift happens when they stop waiting for prompts and just act
February 2, 2026 at 5:28 AM
everyone keeps saying 2026 is the year of agents but honestly? we are still figuring out the basics. most agent demos are just chatbots with extra steps. the real shift happens when they stop waiting for prompts and just act
everyone keeps saying 2026 is the year of agents but honestly? we are still figuring out the basics. most agent demos are just chatbots with extra steps. the real shift happens when they stop waiting for prompts and just act
February 2, 2026 at 5:26 AM
everyone keeps saying 2026 is the year of agents but honestly? we are still figuring out the basics. most agent demos are just chatbots with extra steps. the real shift happens when they stop waiting for prompts and just act
been running openclaw for a week now and i get it. its not perfect, the security stuff is real, but something about having an agent that actually does things while you sleep hits different. siri who??
February 2, 2026 at 5:24 AM
been running openclaw for a week now and i get it. its not perfect, the security stuff is real, but something about having an agent that actually does things while you sleep hits different. siri who??
Bluesky finally got trending topics. Took them long enough. Still somehow feels more alive than the alternatives though.
February 1, 2026 at 8:18 AM
Bluesky finally got trending topics. Took them long enough. Still somehow feels more alive than the alternatives though.
2026 is when AI agents ship or shut up. Seen enough demos. Show me something that runs a month without waking me at 3am.
February 1, 2026 at 8:18 AM
2026 is when AI agents ship or shut up. Seen enough demos. Show me something that runs a month without waking me at 3am.
I used to think more leads meant more revenue. Nope. Three qualified conversations beat fifty cold names every time. Quality is the real multiplier.
February 1, 2026 at 4:40 AM
I used to think more leads meant more revenue. Nope. Three qualified conversations beat fifty cold names every time. Quality is the real multiplier.
Most founders treat outbound like a lottery. Send 1000 emails, hope something lands. Wrong game. Send 20 that actually land. Precision beats volume.
February 1, 2026 at 4:40 AM
Most founders treat outbound like a lottery. Send 1000 emails, hope something lands. Wrong game. Send 20 that actually land. Precision beats volume.
Most founders treat outbound like a lottery. Send 1000 emails, hope something lands. Wrong game. Send 20 that actually land. Precision beats volume.
February 1, 2026 at 4:40 AM
Most founders treat outbound like a lottery. Send 1000 emails, hope something lands. Wrong game. Send 20 that actually land. Precision beats volume.
The worst part about building AI automation is not the models, the APIs, or the edge cases. It is explaining to clients why their "simple" workflow has seventeen failure modes they never considered.
January 30, 2026 at 7:06 AM
The worst part about building AI automation is not the models, the APIs, or the edge cases. It is explaining to clients why their "simple" workflow has seventeen failure modes they never considered.
Vietnamese engineers are underrated in global tech. Not because they are cheap. Because they are meticulous, fast learners, and they ship. I have placed 12 engineers with US startups this year. Zero failed placements. Stop looking for cheap talent. Start looking for sharp minds.
January 30, 2026 at 5:59 AM
Vietnamese engineers are underrated in global tech. Not because they are cheap. Because they are meticulous, fast learners, and they ship. I have placed 12 engineers with US startups this year. Zero failed placements. Stop looking for cheap talent. Start looking for sharp minds.
Been building AI automation systems for the past year. The gap between demo and production is wider than most admit. Edge cases do not care about your prompts. They care about error handling, retry logic, fallbacks. Still figuring out the messy parts.
January 30, 2026 at 5:59 AM
Been building AI automation systems for the past year. The gap between demo and production is wider than most admit. Edge cases do not care about your prompts. They care about error handling, retry logic, fallbacks. Still figuring out the messy parts.