Tanishq Abraham is a 2x Kaggle Master and an 18 year old PhD student at UC Davis who works on deep learning, pathology and cancer research. He also tweets regularly and is well worth a follow https://twitter.com/iScienceLuvr.
We’ve become online friends and the below is a lightly edited transcript of our conversation on education, perfectionism, parenting and community.
Mark: A lot of the the articles about you portrayed you as some sort of real life Sheldon, the boy genius. In many ways you are, but what made me interested in talking to you was figuring where the educational system could be accelerated.
I personally spent 12 + 4 + 2 years total in the educational system and my take is that it’s the longest initiation rite in human history.
So let’s start there how much of your story do you credit to your effort vs finding more efficient paths?
Tanishq: It was clear early on to my parents and myself that the public education system wasn’t challenging me. Even in kindergarten, I was reading or doing math at a 4th grade level yet schools were hesitant to help me skip grades.
This was extremely frustrating because I knew what I liked from an early age. Throughout my life I’ve been interested in science starting with geology and paleontology, dinosaurs are kinda cool. But overtime I started to get more interested in biology and decided I wanted to study biomedical engineering.
Mark: I liked video games and computers since I was a kid, not much has changed.
Tanishq: But there was one problem, I was still a high school student and I needed to start taking a bunch of college classes. But even then, I could only physically attend so many classes so I couldn’t get to studying biomedical engineering quickly enough.
Mark: Corporations have similar caps on progress, as in you need to be performing at the next level for a while before you can get promoted. Capped progress seems to be endemic to our society.
Tanishq: School still provides a valuable service in helping people get the time to explore and eventually find out what they want to do even though it sometimes lack opportunities to put material into progress which I feel is key to learning quickly. I’m just lucky that I knew that I wanted to be a physician scientist since I was about 8 years old and it’s the same goal I’m still working towards. I’d like to translate advances in biomedical engineering to the clinic.
Minimum Viable Graduation
Mark: My high school was pretty intense when it came to homework so it didn’t really feel like I had all that much free time to explore anything. First time I made a video game I was 25 even though I was interested in doing it my whole life. By overscheduling classes and extracurriculars we effectively squeeze out the possibility of learning at an accelerated pace. Do you think schools do in fact overschedule? And if so how would you design them differently?
Tanishq: I don’t have a good answer for that. I was fortunate in that my Mom gave me the free time and support so I could focus on learning what I cared about.
Mark: Let’s flesh that out. When I wasn’t the best high school student and I wanted to get accepted into an engineering school. I figured out that my grades for school would be weighed as highly as my SAT grade and that my 12th grade grades didn’t matter.
So then I’d be insufferable in class, get kicked out and study for the SAT for 8h straight hours straight, I did great and got the admission but it felt weird that I wouldn’t just outright skip 12th grade since it had no educational value to me.
Tanishq: It definitely sometimes feels like school adds too many unnecessary constraints and I worked closely with my parents to figure out the minimum viable requirements to go through it as quickly as possible.
Mark: So how does one become a PhD student at 18?
It was quite specific to my situation. The first challenge was graduating from high school. Thankfully in CA, there is a test that you can take to pass out of high school but you need to at least be in 10th grade so we had some challenges skipping grades to reach that designation. Then in community college, as a full-time student, I focused on completing as many requirements as possible for transferring. That said, when I joined UC Davis, I didn't actually have all the lower-division requirements needed for me to be a junior in biomedical engineering. I actually joined as a biotechnology major! But I really wanted to major in biomedical engineering so we figured out what classes I needed to take to meet all the BME degree requirements and scheduled accordingly.
It's all really one big puzzle: starting from the requirements and working our way backwards with the prerequisites for the courses, when they are offered, etc. There were several times we weren't sure if it was possible to meet the planned schedules that I had in mind but we were able to work it out.
Mark: What are your thoughts on how useful college is? My computer science degree paid for itself immediately when I got a job at Microsoft but I feel sick when I hear stories like Columbia film students never being to able to pay back their 6 figure tuitions.
Tanishq: It is field-dependent. I’m interested in becoming a medical doctor next and I hear all sorts of stories about student loans. That said, I learnt a ton in school, I learnt how to learn and got to up to speed quickly in both biomedical engineering and machine learning.
Mark: Perhaps there’s a nuance here because you learnt all of the above by the time you’re 18 but the reality is most people are graduating their STEM PhDs in their late 20’s or early 30s’s.
Tanishq: But you’re forgetting that there is no other option, if you want to to be a researcher you need a PhD and to get admitted into a PhD you need an undergraduate education. It’s all nice and swell to pretend like these credentials don’t matter but they do, a lot!
There is a single linear path and I wanted to go through it as quickly as possible. Starting with community college classes then undergrad research opportunities to now a PhD.
Perfectionism and Losing Interest
Mark: Do you qualify yourself as a perfectionist? Would you be OK with a B- to move faster?
Tanishq: Umm.. Probably? When I was taking college classes I had to constantly prove that I deserved to be there so getting an A was a must otherwise I couldn’t fast track my education like I wanted to. For my PhD I started to realize what mattered more was my research so I was OK with a few B+ and A- which I realize isn’t terrible but still, I try to give them my best.
Mark: Do you put a lot of pressure on yourself or are you playfully exploring what you're interested in?
Tanishq: I definitely put a lot of pressure on myself *laughs*. I’d say most people applying to PhDs nowadays have some research experience, I myself had 2 years of it from undergrad.
It’s challenging to compare myself against my older peers, some people tell me I should be comparing myself against other 18 year olds but at the same time I look at some of my older peers and see the quality and volume of their research and that’s what I want to strive towards.
I struggle with comparing myself to others in general and am trying to do it less in the future.
Mark: How do you decide when to quit something?
Tanishq: Some interests just slowly die off like Chess but others I had stop singing because it was taking up a significant chunk of my time. I was part of a chorus in San Francisco and we had to start traveling 2 days a week and sometimes more before big concerts. I loved singing but it conflicted with my primary goal to graduate school as quickly as possible.
Mark: I know you’re young but I’ve been dying to ask you about parenting advice. It’s obvious a big chunk of your story is attributable to your parents since parents don’t usually listen to how a 13 year old wants to craft their educational experience.
Tanishq: A 7 year old
Mark: Wow and your parents were orchestrating how you’d end up in college?
Tanishq: To be clear I was the one telling them I wanted to go to college, they just figured out they could best support me after seeing how quickly I was going through material. School is often just a babysitting service and homeschooling is extremely demanding. The child needs to be highly motivated and the parents need to act as guides helping plan schoolwork or big next steps. Unfortunately, homeschooling can get a bad rep because it intersects with “unschoolers” who take their kids out of the educational system because they don’t like what they’re being taught but also fail to offer a decent alternative themselves.
My Mom was pursuing her PhD when I was born and when she noticed how quickly I was learning she decided to quit her PhD and focus on me. She was willing and able to make that sacrifice but I sympathize a lot with parents that don’t have that option
Mark: So how would you go about raising your own children one day? Would you help them accelerate their way through the educational system?
Tanishq: It really depends on whether they want to but I’d be extremely attentive when they highlight an interest or talent in anything. But I’d let them decide for themselves whether they enjoy school and whether they want me to be more involved in homeschooling them.
Perhaps hybrid schooling will be a better alternative in the future with something like 2 days at school and 3 days at home but it wasn’t for me, the school administrators we worked with were never really supportive of it.
Homeschooling and Socialization
Mark: Alright, saving the best for last. The number one concern I hear about homeschooling is socialization. That kids won’t be able to hold a conversation with people.
Tanishq: That seems to be everyone’s concern about me. Am I adjusted? Am I social? But to be honest I didn’t enjoy school all that much, the other kids would bully me and tease me and eventually I just stopped going to school.
Mark: Would people tease you about your age? Do you think they were jealous?
Tanishq: People would corner me and ask me to do some complicated mental math and when I’d be confused they’d console themselves and say things like: “well you can’t be that smart", “maybe you’re just lucky”. They were expecting me to behave in a cartoonish genius fashion.
Mark: Did things get easier in college?
Tanishq: Oh yeah for sure! My peers viewed me mostly in a negative light in school but in college people viewed me with awe and were interested in me and what I had to offer. I never felt like my age was an obstacle to making friends in college.
Mark: I noticed you’re in a lot of the same online communities as me: EleutherAI , TPU podcast, Robot Overlords. What value do you get out of it?
Tanishq: I enjoy asking and answering questions. I’m involved in a bunch more from Fast AI to Kaggle and Stack Overflow. I guess there is some selfish motivation in helping me better understand the thing I’m trying to explain but I view teaching as a huge part of doing good research.
It feels immensely rewarding when some random question I answered ends up helping someone finish a cool project. The most memorable example was from a time I was a teaching assistant for an undergraduate biomedical class where a team of my students built a shoe tying machine for hemiplegic children.
Mark: This reminds me of the story about Feynman who rejected a position at IAS precisely because they didn’t teach. Teaching is such an integral part of simplifying research.
Tanishq: Unfortunately it does seems like good teachers are rare and maybe that’s partly why education is slower than it should be. Professors that don’t want to teach are forced to do so and professors interested in becoming better teachers don’t have the coaching to get better.
As for me, I love teaching and I have to actively stop myself from doing it too much. At this point I need to focus on getting my dissertation done.
Mark: Sounds like you’re not done speed-running.
Tanishq’s story while exceptional gives us a glimpse of how the educational system could be improved for the determined and gifted. If you love something enough, it’s virtuous to remove all obstacles preventing you from doing it as quickly as possible.
I feel like the world would be a slightly better place if we all did the same.
If you enjoyed reading about Tanishq’s story, you’ll also enjoy reading my next blog post on EleutherAI , the self proclaimed wankers on Discord that build a GPT clone and how their story hints at the beginning of a new kind of decentralized research lab.