My 1st Panic Attack

On July 17th I built and launched the MVP of LearnFromAnyone.

On July 20th I quit my job to work on it full-time.

Nervous doesn’t even begin to describe how I felt that morning. Knowing that I was going to have to tell my team that I was leaving was tough. It was an emotional day for me.

I love Hivewire. I loved the work, I loved the product, and most importantly, I loved the people. Getting to work with that team for a year is something I wouldn’t trade for anything. Showing up to work every day felt like waking up and going to hang out with my friends. They gave me my 1st opportunity and invested a lot of time and energy in helping me grow. Not a day goes by that I don’t feel immense gratitude for that experience and for those people.

The night before I quit I didn’t sleep. At 8:30am, I drove across the street to our office. After about 10min of sitting in the building, I told our CTO who I reported to about my plans. I let him know that while I loved working at Hivewire, that this was something I had thought long and hard about and that it was my final decision. It was something I had to do. If I didn’t do it I’d regret it for the rest of my life.

He let the rest of the executive team know. Our CEO came back to where I was waiting, and I’ll never forget what he did. He came around the corner beaming, put his arm around me, and looked me in the eye and said “Man, I’m so pumped for you!” That’s leadership. That’s how you empower people.

He brought me in with our COO and CTO and we chatted. They even made sure I knew I was still invited to our work party that was in 2 days (Covid safe movie theater rental - it was awesome). We all got on the same page, and then they gave me the opportunity to say goodbye to the rest of the team. It was hard - the definition of bittersweet. But 90 minutes after entering the building I left with my belongings. I was officially in full startup mode.

Things were great.

I went home and immediately started working on product. I was hyped. There was no way I was going to let that energy go to waste. I was just waiting on the green light from OpenAI to re-launch LearnFromAnyone following my security review, and until then, I was going to keep improving it.

But at 3pm I hit a wall. I’d slept maybe 12 hours over the last 3-4 days. I was sleep-deprived, so I decided to take a nap.

I woke up to a message from OpenAI. I was not cleared to launch.

I was devastated. I’d been telling people - thousands of people - that we’d be back up in a few days. But OpenAI wanted to review a few more things and then meet later that week.

Perfect. Just perfect. I had become so caught up in my own excitement that I failed to consider that this was the likely outcome. Now I had to tell all of these people I’d been hyping up about the bad news. That’s when the negative thoughts started to pile on.

I had no job, no income, and no live product.

All I had was the runway I had in my personal savings and a rough, early product that I’d maybe be able to re-launch.

Imposter syndrome kicked in. Anybody that knows me will tell you that I’m about as cool and collected as they come. Imposter syndrome is something that’s never affected me. I’m a pretty relentless and generally unphased person, so suddenly having to deal with intense self-doubt was something I was not familiar with.

I started to sweat. I couldn’t think. I was feeling extreme fear. Then the shaking started. I was having a full-blown panic attack - something I had never experienced. My mind was rattling off about how I was an idiot for quitting my job and that I had absolutely no idea how to build and run a company. This intensity lasted for about 20 minutes - it was scary.

Eventually I decided to head outside and go for a walk. Walking always helped me clear my head, so I walked, and I was out for almost 2 hours. Over the course of those 2 hours I was able to gradually calm down and compose myself. I told myself that it was completely natural to feel such a rush of emotion on the day I made a massive life change - especially considering my severe lack of sleep.

By the time I got home, I was okay again. By bedtime, I was 100% hyped up again.

I tell this story because I think it’s important to talk about our challenges. It’s important that we normalize talking about pain and fear and uncertainty - those emotions are part of being human. There’s an unspoken expectation that people act tough and like they have no problems. This hurts everyone.

Starting a company is a big leap. So are a lot of things. Feeling unqualified to do something new is normal. There’s nothing wrong with you if you go through phases of feeling overwhelmed and unsure of yourself.

But you’ve got to address your problems head on if you want to overcome them. Push through the pain and face your fears.

3 weeks later I’m in a great spot.

We got cleared for launch. LearnFromAnyone is wrapping up our 50 user alpha that has been a resounding success. We’re doing everything we can to continue to build and iterate on product while managing our expectations as we work towards permission to onboard more users.

I’m more excited and optimistic than ever. I’ve found a routine that works well for me and a support system that I can lean on as I learn how to build and run a company. Startup life is good.

I anticipate that I’ll have more thoughts of self-doubt. Maybe not as severe as the 1st wave, but inevitably they’ll come back from time-to-time. And that’s okay.

Self-doubt is normal, and we all have the capacity to overcome it.

We’re far more talented and capable than we let ourselves believe.

The 2 Sigma Problem

In 1984, educational psychologist Benjamin Bloom published a paper called “The 2 Sigma Problem.”

The paper highlights an important discovery. Students tutored 1-on-1 using mastery based learning performed two standard deviations higher than students who learned through typical methods. Put simply in Bloom’s words, “the average tutored student was above 98% of the students in the control class.” Further, Bloom and his research team found consistency in results from students in all grades and in all environments.

This tells us a few important things:

1) Our system is setting up students to be average.
2) All students are capable of mastery.
3) Students need personal instruction to thrive.

Let’s dig into each of these three ideas.

Our current education system sets students up to be average. It’s diametrically opposed to what Bloom found to be effective. Rather than using mastery based learning and 1-to-1 tutoring, our system consists of massive class sizes that spread teachers thin and awful grading standards that do nothing to make sure students actually learn and retain knowledge.

This is a massive problem.

Students fall out of love with learning and natural curiosity at an early age thanks to a system that better resembles quasi-educational daycares than effective learning and curiosity incubators. We can do better.

In 1968 Bloom proposed “learning for mastery” - a method where students must demonstrate a given level of mastery on a part of a subject before moving onto the next part. Rather than giving all students the same amount of time to learn a subject, each student gets as much time as they need to become proficient in the material. Our current model resembles memorization and assessment - not learning and teaching - and mastery based learning can help fix that.

Check out these two TED talks by Khan Academy founder Salman Khan for a great explanation and demonstration of how effective teaching for mastery is:

1) Let's teach for mastery -- not test scores.
2) Let's use video to reinvent education.

While mastery learning is a great step, we also need to address the problem of large class sizes and a lack of personal instruction.

We put both teachers and students in a horrible position. More students and less funding means that class sizes are larger than ever. Teachers try to make do with the resources they have, but they’re put into an impossible situation - 1 teacher isn’t enough for 30+ students. They’re spread too thin. And now they also have to juggle remote learning.

In video #2 above, Sal Khan talks about how video instruction helps address this problem. Recorded lessons allow students to pause, rewind, and rewatch the material allowing them to learn at a pace that’s natural to them at a time that’s convenient for them. Teachers can effectively record their lessons so that students can get a more personalized experience despite the lack of direct attention from the teacher. Again, this is a step forward, but true 1-on-1 tutoring requires interaction.

At the time, Bloom concluded that 1-on-1 tutoring was "too costly for most societies to bear on a large scale." He raises an important point. Getting every student a personal tutor is a burden than society can’t bear. We have neither the number of teachers needed nor the resources to fund them. However, technology will be able to solve this problem, and it will be able to solve it much sooner than we think.

Advances in AI and AR/VR technology will soon allow us to create virtual tutors for our students. Not only will this eliminate the need for human tutors, but these tutors will be far more effective than human tutors ever could be.

We’re fast approaching a world where every student will have a customizable, personalized tutor that is an expert in every subject in their pocket.

A world that looks like this:

This is the problem I’m working on at LearnFromAnyone. In front of us lies the path to building the scalable solution to Bloom’s 2 Sigma Problem, and LFA aspires to be the solution to that problem.

While implementing one of either mastery based learning or 1-on-1 tutoring will raise students by one standard deviation (84th percentile), implementing both of them will raise them by two standard deviations (98th percentile).

A future where tomorrow’s average student is equivalent to today’s 98th percentile student is a future worth building and a future worth aspiring to.