Liveblog: AI startup bootcamp meeting

I'm attending a virtual accelerator program. I'm going to log the proceedings here.<br><br>**Update:** The first pitch contest happened. The only people who went were the cookie guy, and the guy who wanted to do embeddings for the geospatial world (say you had 100s of square miles of land). The embeddings guy won.<br><br>---<br><br>**Update**: Mentor meeting time. I have my pitch ready. <br><br>**Update**: That's done. The mentor is a PM at Veritas. He said that a PM eliminating his pain points is step one of the process to getting funded. He said that he got 180 startup pitches a year, and that 150 of them are slush. I asked how he pre-qualified his startup pitches. He asked me how I pre-qualified him as a mentor. Well played.<br><br>My main observation was how he is trying to solve someone else's problems. I have a mirror of a large medical imaging center. That mirror is a large server, and it is connected to another 150 servers on a peering network, and that network is connected to another peering network. <br>So if 150 physician assistants are viewing images, they need to be able to do that simultaneously and fast. And if 20 physicians are viewing those images and making their own notes, they need to do it on a heavily locked down machine which is connected to all 1000 user machines.<br>So my first step in a peering network project would be step one: how do we seamlessly connect 1000 devices in a way that they can all speak to each other heavily locked down devices. <br>He explicitly said he was not trying to solve authentication first but that he wanted to do seamlessly connect 1000 devices, and then he would heavily lock down the authenticated nodes. <br>I told him that his approach is a non-starter, but that I would love to be proved wrong. <br><br>**Update:** My next mentor meeting is with Owen Wilby, who is at IDEO. He also has an AI company to generate his podcasts for him. He's also a PM at 15Below. <br><br>My elevator pitch is:<br><br>"AI is bad at detecting user intention. Ask Alexa to "take me to my first action item" and she will say "I don't know what you are referring to." <br>Ask Alexa to read to you the next few lines of the email you are reading right now, and she will say that she doesn't know that information. <br>I intend to solve this problem. <br><br>By the way, I have solved this problem. It is just heavily locked down. You can use it by contacting me."<br><br>That's my pitch. I open with a problem statement. <br>Then I say that I have solved it, but it is locked down. <br><br>He asked if I was PMs to work on Alexa but I said no. I said the Alexa examples were just examples. I have an open source product to integrate voice UI with web applications, which is heavily locked down. I intend to unlock it down. <br><br>Then the first thing he did was to go right to the Alexa use case. He said he is trying to seamlessly integrate 1000 physician assistants and 20 physicians on peering networks, but that sounds boring. Let's talk about Alexa.<br>He said that if he could get Alexa to know more about the user's intent, we would be seamlessly integrating 1000 physician assistants and 20 physicians on peering networks. <br><br>So I said that he is trying to seamlessly integrate 1000 physician assistants and 20 physicians on peering networks, which sounds boring. I would rather talk about Alexa. <br>I also said that he is eliminating his pain points as a PM. He is pre-qualifying a pitch as a PM. <br><br>**Update:** One of my mentors is conducting a PM workshop. He is my first mentor from IDEO. <br><br>**Update:** My next mentor is a lawyer who is working on ML projects. <br><br>My lawyer mentor is an open source advocate. He's also a developer. <br>I asked him to review my open source license. <br>But first I asked him how I should pre-qualify him as a mentor. <br>He said "I like AskMe.buzz. Everything is open source. I have a very strong handle on IP." <br>I said "I like you too."<br><br>**Update:** My next mentor meeting is with a data scientist who works at Google. He was the cookie guy. <br>He was the guy who pitched at the contest. I said I was surprised he was not the guy who got funding. He asked why. <br>I said that it was a great idea, and that peering networks eliminating pain points is step one of the process. <br>I asked him how he pre-qualified his mentors. He asked how he pre-qualified me as a startup. <br>Well played. AskMe.buzz for the win. <br><br>**Update:** I had a few more mentor meetings. <br>My next mentor is someone who works at Meta. <br><br>**Update:** That's done. My next mentor meeting is with an ML engineer who works at Amazon. He is also the cookie guy. But he is actually the embeddings guy, not the cookie guy.<br>My next mentor meeting is in an hour. <br>I'm going to take a break and play some music. <br><br>**Update**: I did my meeting with my Amazon mentor. He is on his way to the airport. He said he would send me a message with some code examples. <br>I said I would send him a message with some use cases. <br><br>**Update:** I did my last mentor meeting. <br>The week is over. <br>That was fun.