- Fast Principles
- Posts
- Meet my new personal AI knowledge assistant
Meet my new personal AI knowledge assistant
Meet a new AI agent platform enabling 24/7 advice for human experts

Inside Issue #24
Introducing Jason jr. an AI knowledge assistant via Tmpt.me
I interview tmpt.me’s CEO and co-founder Scott Zimmer
This week’s AI magnet is all about regulation, policy and politics
Introducing a New AI-Powered Knowledge Assistant
In today's fast-paced digital age, finding reliable, accurate, and personalized information can be a challenge. That's where Tmpt.me comes in. The new AI-powered knowledge transfer platform empowers Experts, the thought leaders on the platform, to share their knowledge with the world, creating digital twins that provide personalized guidance and support to knowledge seekers.
I asked the Tmpt.me CEO’s knowledge assistant to tell me what it is and it replied with:
Tmpt.me is a platform designed to expand the knowledge-sharing capacity of Experts, helping them reduce burnout and multiply their impact.
Introducing Jason Jr.
To demonstrate the power of Tmpt.me, I've created a digital twin of myself, affectionately named Jason Jr. My wife tells me this is terrible branding. She’s right, but I feel it’s on par with other poor brand names like Claude and o1 to name a few…
This isn’t just basic search and recall of a search engine, but rather uses my customized knowledge base, my weights of key content and customizes the responses to users of the agent to cater to their specific line of questioning. It will even cater to stylistic word choice by the user to support comprehension.
The Importance of a human-centered Data strategy
The quality of an AI agent is directly tied to the quality of the data it's trained on. That's why I've carefully selected my past work to ensure that Jason Jr. is well-equipped to provide accurate and helpful information. Experts on Tmpt.me have a dashboard that makes it easy to continually add content and occasionally update their agent’s answers with revisions that delight their users. This positions the Expert at the center of an ongoing feedback loop.
The future of AI-powered knowledge assistants
I'm excited to use Jason Jr. to support my work and answer questions 24/7. It’s already been instrumental in answering questions about my recent Action Chain Workbook on strategically planning AI agents. As I continue to develop my expertise, Jason Jr. will upskill alongside me, providing even more valuable assistance to my community.

A history view from my Expert dashboard:
I have the the ability to refine any response shared on my behalf, resending the message and updating my knowledge base for future inquiries.
The importance of a human-centered data strategy
Tmpt.me is constantly evolving. The assistant builder makes it easy to add new content, and will even regularly crawl sites that are starred for updates like Fast Principles. If someone asks a question the teacher’s assistant hasn’t seen before, it doesn’t hallucinate it will state it is a good question. and will alert Jason (the author) and he will follow-up shortly. An alert is triggered and the question can be responded to which updates the chat with the user and updates the knowledge base.

A content view from my Expert dashboard: As I upload new content, I confirm the label for source attribution.
The Future of AI-Powered Assistants
Tmpt.me's human-centered approach sets it apart from other AI agent builders. By focusing on creating digital twins of real people, Tmpt.me empowers individuals to share their knowledge and expertise in a more personalized and effective way tuned to knowledge transfer and upskilling of human users.
I'm excited to share Jason Jr. to support my work and help others. It’s's already been instrumental in answering questions about my recent Action Chain Workbook on how to plan and build ai agents. As I continue to develop my expertise, Jason Jr. will grow alongside me, providing even more valuable assistance to my community.
Q&A with Scott Zimmer, Tmpt.me CEO

1. Why did you start Tmpt.me?
The origin story for my co-founder Andy and I tracks back to our experience as senior leaders of innovation and technology teams where we learned that investing in helping people grow is one of the best things you can do to create a high-performing team. We were always experimenting with different methods of how to help people, and the recipe that became our go-to was to pair the growth seeker with an expert who was willing to offer their guidance over time. This would typically work great, unless the expert was widely revered, in which case they’d have little or no capacity to assist.
So, the opportunity space that inspired Tmpt.me: the people we most want to learn from are also those with the least capacity to offer.
2. What is the unique value proposition of Tmpt.me compared to other AI-powered tools?
Our unique value proposition is trust, which we deliver via human-centered AI.
We all have the ability to learn from google, ChatGPT, etc. but we’re left wondering “can I trust what it’s telling me?” and “whatever’s behind this, do they/does it share my values?”. We’re facing this right now as a nation with TikTok, wondering if we can trust the algo.
So we purposely position Tmpt.me as human-centered AI and we obsess over making it easy for Experts, our creators, to build Knowledge Hubs where they have transparency and control over the way their knowledge is shared, so it can be trusted by all.
3. What is the difference between a human-centered AI assistant and a digital twin?
Honestly, when those alternative names popped up, we didn’t imagine there would be a big difference. However, what we’ve learned is that a typical digital twin or AI clone is something that is trained once and unleashed to do its job, often with very brittle instructions for a specific context. In contrast, we’ve put a ton of energy into ensuring Experts on Tmpt.me are positioned at the center of a feedback loop, where they see any answer their AI Agent generates, and they can easily edit it which in turns enhances future interactions. With this approach, we extend an Expert’s reach and deliver their knowledge in context, but keep them actively engaged with minimal effort.
Sure,I had a lightbulb moment many months back when I offered my Calendly link to someone who was looking to meet with me: it occurred to me that they didn’t really want my time, they wanted to get my input on a few key topics where they respected my expertise. In fact, they needed the input right away. but would reluctantly wait until my next available meeting slot. So I updated the way I’m responding to inquiries: if you want an answer right away: use my Tmpt.me link, if you’re ok waiting, here’s my Calendly. On our Tmpt.me website, I just go straight to offering my Help Agent on the bottom right so that anyone who might be Tmpt’d to book a meeting can get the easy Qs answered first.
5. What are the biggest challenges you've faced in developing Tmpt.me?
Our human-centered approach has really worked, but it’s a totally new paradigm for people. Up until now, there has only been static published works or live interaction with humans. Now live conversation with the AI Agent representing the human feels so unfamiliar that many just choose to stay with what they know, with all its limitations.
Sequoia has a great 3-part PMF framework where the third category “Future Vision” fits us perfectly. As with much of the AI space, when you’re making sci-fi a reality, you need lots of great examples of renowned people using the new service and getting benefit from it. Lennybot.com is a great example that I like to point to, powered by another startup in our space, and MasterClass adding the AI on-call feature using a teaching assistant representing the expert is an exciting mainstream use case.
6. What are your plans for the future of Tmpt.me?
Right now, we’re most excited to continue to onboard more Experts who are Tmpt’d to share what they know at greater scale.
One feature launching soon that I’ll name specifically is “conversation insights”. This new feature will capture a conversation, distill the insights that would make great additions to the knowledge base, and add them automatically, all while ensuring the Expert has full control. It’s a whole new frontier of knowledge capture that will enhance the value of our AI Agents tremendously.
Looking ahead, we’re excited about the way Multi-Agent Systems (MAS) will enable a new form of human-ai collaboration, one based on hyper-personalized, just in time applications and agent systems that empower all software, documents and media to be tailored specifically to the tastes of human managers.
7. What is an example of a content creator using tmpt me in their business?
There’s a reason 400,000 professionals read this daily.
Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.
🧲 AI Agent Magnet
A new paper authored by Google that defines agents and agentic systems architectures to align these systems capabilities. They highlight these agents can use tools, make decisions autonomously and execute tasks in a team — like a capable assistant or teammate. This paper is important as builders and policy-makers must all align on the definition and architecture of these systems to appropriately assess risk and create effective guardrails. This paper is a good companion with the paper I co-authored as an agent primer.
This combination of reasoning, logic, and access to external information that are all connected to a Generative AI model invokes the concept of an agent.
What will become of the derivative AI generated work as it relates to copywrite infringement or protection? Today, this is a mess — no country has ruled on it, but this consultation review does give us clue. The current preferred summary in this approach:
Remove specific protection for computer-generated works. This is the government’s preferred option, as there is no evidence that the presence of copyright protection for outputs has economic benefit, and on the contrary, there is strong opposition from some creators.
|
Sharing another newsletter, Zain Khan’s Superhuman is one of the largest AI newsletters focused on productivity hacks using AI agents. If this sounds useful, feel free to subscribe — I do.
💬 Suggestion box
A newsletter exploring the principles that will help us design AI agent experiences and startups that amplify human creativity.
Subscribe to join a community of designers and developers shaping purposeful AI agents.
How'd we do this week? |
New Referral Rewards
Refer a friend and get our latest digital download.
Until next time, keep innovating and stay curious!
PS — A huge thank you to Scott Zimmer and Andy Jenkins for your generosity of your time and support for this week’s issue with the demo on tmpt.me.