Tyson Thomas
@tysonthomas90
Followers
164
Following
264
Media
22
Statuses
406
Engineer & Entrepreneur tackling complex engineering puzzles. Building the open-source AI browser @browseroperator. Innovating at the edge of tech.
Seattle, WA
Joined May 2010
What if your browser AI agent could help you find influencers on X! . What if your browser AI agent could find you the most trending posts that you care about!. What if you could view a version of X that's custom to you! . Now @BrowserOperator can! #AIAgent #operator
2
0
6
It's the reward signal š.
Entrepreneurship rewires your brain and this is how. You have a startup idea. You ship it. Most ignore it. You keep going. A few strangers swipe their credit cards. You iterate. You figure out how to get more attention. More customers sign up. They stick around longer. They.
0
0
0
RT @garrytan: Hire the engineer who worked on the best product. Hire the marketer and sales person who managed to get the worst product toā¦.
0
163
0
At @BrowserOperator we convert web pages to structured data so that LLMs are clear about its contents!.
You know those AI apps that always work? They all use JSON prompts. Started using the same format for my AI prompts and my outputs became shockingly consistent. Here's exactly how to write JSON prompts (with code, screenshots, real examples):.
0
1
1
LLM providers should make OAuth Connect the default way to simplify app connections and streamline BYOK!. @OpenAI @AnthropicAI @grok @Kimi_Moonshot @GroqInc.
1
0
0
Its awesome that there is OAuth Support by @OpenRouterAI! Help's simplify setup for users! Let's go!!.
š We have a major release - just in time for the weekend! You can now connect to your @OpenRouterAI account with one click, getting started in seconds!
1
0
1
Open source AI is gonna lead more adoption across industries!.
Remarkable how far we've come. From fear mongering OS AI across VC, academia, AI labs, and politicians to full throated endorsement. Thank you to everyone who has taken a stand on this over the last couple of years. In no small way have you helped sway and save a nation.
0
0
0
RT @AnthropicAI: The White House AI Action Plan gets it right on infrastructure, federal adoption, and safety coordination. It reflects manā¦.
0
67
0
RT @sriramk: šŗšø. Today is a day we have been working towards for six months. We are announcing Americaās AI action plan putting us on the rā¦.
0
799
0
RT @AnjneyMidha: The new AI Action Plan from the @WhiteHouse calls for the US govt to 'create a supportive environment for open models'. Itā¦.
0
77
0
RT @TheDaveClark: This is Michael Truell, CEO of Cursor. He's the founder behind the fastest-growing AI startup in history ($0 to $300M ARā¦.
0
31
0
Only a few spots remain for the SeattleJS July Workshop: "How to Build Tiny Agents in JS" with Tyson Thomas!. Learn AI agent architecture, effective prompts, tool integration, and web deployment. Ideal for developers of all levels. Secure your spot now:
lu.ma
š Get ready to learn and connect at SeattleJS! š Prepare for an evening filled with fun, networking, and learning. Whether you're a seasoned developer orā¦
1
0
1
RT @vkhosla: So true. Windsurf and others are really bad examples of founders leaving their teams behind and not even sharing the proceedsā¦.
0
47
0
RT @StartupArchive_: Y Combinator CEO Garry Tanās advice for startups: āWhen youāre small, act smallā. A lot of founders try to emulate larā¦.
0
102
0
Imagine the possibilities if we had 100x compute on our laptops!.
NVIDIA CEO: āI have computing anxiety if I donāt have enough computersā . > You know at nvidia, we do A LOT.> chips, systems, networking, algorithms, software .> none of it is possible without AI now.> so we built ourselves many supercomputers for our engineers.> but guess what.>
0
0
0
Very interesting read on context engineering!.
After four overhauls and millions of real-world sessions, here are the lessons we learned about context engineering for AI agents:
0
0
1
š Breaking: Unlock the hottest AI models INSTANTLY upon release! Dive into @Kimi_Moonshot K2 magic powered by @GroqInc on @BrowserOperator no waiting, pure innovation! Who's ready to level up? š„ #AIPower #FutureTech
0
1
0
RT @exolabs: We're doubling the number of Apple Silicon macs that can train together coherently every 2 months. Our new KPOP optimizer wasā¦.
0
20
0
I'm interested in how others have addressed scaling agentic workflows. Large AI labs use varied approaches, such as multi-agent systems, agent handoff patterns, or Devin's method of passing context to all sub-agents.
cognition.ai
Frameworks for LLM Agents have been surprisingly disappointing. I want to offer some principles for building agents based on our own trial & error, and explain why some tempting ideas are actually...
@Kybervaul @AravSrinivas Good question @Kybervaul! Smaller, focused tasks help agents solve problems better. Many large agent frameworks have shifted to multi-agent systems for this reason. Here's an interesting read by Anthropic:
0
2
2