I just got my @EFDevcon ticket — paid for with ETH!
Next stop: Mumbai 🇮🇳 Join me at Devcon 8 from November 3–6, 2026 for four days of big ideas, technical depth, community, and the people building the future of open source technology.
devcon.org/ticket/Ujjwal/msq…
Ujjwal retweeted
It's possible to get high on math.
Here's how.
3 steps.
1. Work problems until your brain is mush
2. Watch a math video that's out of your depth but has pretty animations
3. Philosophize, try to make sense of what you just saw
You won't have the background for a proper understanding, and your working memory will be too taxed to keep your logic on the rails, and your brain mush will make for a good "association stew," so you'll end up trying to bridge the gap in your understanding by making weird associations.
You can do this with other fields too.
But realize that steps 2 and 3 are not productive to learning. You can't count them towards your actual study time.
Ujjwal retweeted
Feeling drained doesn't automatically mean you're doing too much.
Sometimes it's because you're never fully doing anything.
You work while thinking about resting.
You rest while feeling guilty about working.
You're always halfway somewhere else.
And halfway is exhausting.
When it's time to work, work.
Immerse yourself. Give it everything.
When it's time to rest, rest.
Actually recover.
When you go to bed, actually focus on sleeping.
Don't turn rest into a half-conscious state of worry.
Don't half-ass your rest.
If you're not full-assing your rest, then you're not full-assing everything.
Applying this to zk, i would say, go build a strong maths foundation first, calculus(for mathematical maturity) , linear algebra, abstract algebra, discrete math, intro to probability and statistics, programming and data structures.
And then you will have the foundational skill to go to actual zk stuff, and then master it.
go to cryptohack, solve a lot of ctfs to get the good intuition on cryptography security.
At the end of the day, whatever field you wanna choose, strong foundational skill, and that takes time, and "using ai won't reduce that time to build the foundational skill, it can accelerate by say 3-4% at max" .
Building strong foundational skills will always provide a massive edge. Most people just don't want to put in the work and will grasp at any excuse not to do so.
If you are choosing not to develop serious technical chops now because of AI, then don't kid yourself, you were never actually going to do it in the first place, you were always just going to grasp at the nearest excuse, and AI is the excuse you landed on.
Go listen to "Because I Got High" by Afroman and replace the word "high" with "AI". That's you.
I think there should not be any maths light introduction to zk, there should he hard pre-requisites on basic number theory, linear algebra, and abstract algebra.
It creates a false idea that its maths light, which it is not.
And also you will fail to see the elegence of zk.
Ujjwal retweeted
#1 killer of creativity: lack of foundational knowledge/skills.
#1 unlock for creativity: having information subconsciously accessible and effortlessly manipulable in your brain.
Your brain wants to find patterns and connect the dots, but it can only do that if
1) all the data is available for background processing, and
2) manipulating that information takes so little effort that it doesn't derail your big-picture train of thought.
true. Especially in mathematics, i have noticed that if your fundamentals are really down, the other concepts based on them will start to make sense much quicker than compared to when you have little experience in the pre-requisite knowledge and problem solving.
And the best way to fit that knowledge in your head?
1. practice - yes, as long as you dont practice enough number of problems, it wont stick to your head.
2. spaced repetition - if you solve many problems in a single day, and dont touch the concept for weeks, you are bound to forget most of it. So, instead of solving all the problems in a single day, space that out properly.
But how much time gap to take? initially few days, then few weeks, and finally after few months.
The people who actually competes in IMO, arent the ones who only practised only the diffcult IMO pyqs, rather they practice so many easy problems, and ultimately that builds the intuition to solve hard problems.
And this is not just limited to mathematics, it also applies to programming, and other subjects as well.
The single most important word in the science of learning:
"automaticity" - the ability to carry out basic skills effortlessly without conscious thought.
Without automaticity, you have to consciously think about every low-level action, which overloads your working memory and leaves no room for higher-level reasoning.
Okay, so I am going to document my current ongoing(feb-now) journey to learn cryptography and ZK .
I actually came to know about zk in dec 2025.
I was so fascinated that i thought lets do it, only to realise few weeks later what have i taken a dip into.
And from mar 2026 , Moonmath Manual study group started. Initially , since no one was coming forward to lead the discussion, i put myself out to lead the sessions. Few days later Nicolas Huber(PhD scholar) also joined to lead.
Why this ?
1. accountability
2. better articulation of thoughts - I(we) mostly think, we know something until the time comes to actually explain it to someone. And I(we) freeze.
3. inspire other ppl.
The format is gonna be very popular 1001 from @0xLoopTheory
Although i have started my journey in feb 2026, but i am starting this documenting stuff , after almost 5 months.
We are still going through @LeastAuthority moonmath manual, so if you are interested, you can join us by joining
Grateful for the community :
@__zkhack__ for the platform, whiteboard session, etc...
Nicholas Huber - for leading the sessions with me.
@AnnaRRose - for her consistent contribution in zk community, manages podcast, zkmesh, zkhack, etc
@LauriPelto - giving insights
@LeastAuthority for the moonmath manual
@Jeyffre he is the reason i came to know about zk
@RareSkills_io - for the zk book
We are currently going through @LeastAuthority moonmath manual chapter 6.
We have went through arithmetic, abstract algebra, and elliptic curve. And it was a pure ride.
If you wanna join, join the zkhack discord.
The discussions are lead mostly by Nicolas Huber and me
I know i(we) have a long way to go. But why not enjoy the process.
Cheers...
@jpmorais80 hey, i saw your reply , but couldn't read completely, which i think you deleted. it was something on moonmath manual.
i would appreciate your feedback.
So true, we regret some of our past choices. But maybe thats life, combination of some good choices and some bad.
And the realisation that, "its okay", we can't change our past. And learn to live with it.
Even, i have experienced this(multiple times):
1. if i prepared for my dream college entrance, isi b.stat instead of trying to find some "quick cash" scheme, i would be in much better position that i am in now.
2. if i would have prepared whole heartedly for maths olympiad during my high school, i would be in much better position than i am in now.
3. if i wouldn't have quit iit/isi prep, i would be in much better position than i am in now.
4. and many more...
We accept that we made mistake, and move on with some learning.
When I was doing my Master's in Physics, a group came to our university to offer a PhD scholarship in oil physics. Half of it would be done in Brazil and the other half in France.
I humbly declined, eager to pursue my PhD in astrophysics and cosmology.
It was great to learn about black holes, D-branes, primordial density perturbations... but I'd be in a much better financial situation today if I had chosen oil physics. Money is there, not in black holes.
When we're young, we're so idealistic! When we're old, we got tired of being Elizabeth and just want to be Mr. Darcy.
kudos to the people who are participating in protest.
We youth are fighting for our own (and upcoming) generations. Wish i could also participate.
Really appreciate all the efforts.
There comes a time in your reading journey that you start drifting away from non-fiction books, and starts to learn(and enjoy) a lot more from fiction books, like classic novels, thrillers, thrillers, etc
If a person have x thousands (even millions) follower , it doesn't make them the source of ultimate truth. They might have thier own agenda behind whatever they are saying, tweeting, or promoting.
Use your own critical thinking skills.
Very true. I see people even commenting using ai, writing bio, posts, all using AI.
Can't you even form a sentence in your head. If you make some grammatical inconsistency, that's okay. I dont want to talk to a robot.
Again, dont outsource your thinking, so simple.
I don't get why a lot(maybe most) guys are pro ai, as "if you aren't token maxxing, you are missing out a lot".
I don't think so. I use AI very little, and only when i need it.
I see even AI as a "tool" to accomplish things, and not be over relied on this.
The actual reality is that many security researchers still are not AI-first in their work. It's a disadvantage. It should be obvious to you already, unless you are a big dummy.
To start with AI and truly "believe" in it, use open source Skills. Read & study the SKILL.md files. Try to improve them, fork them, build on top of them. Learn by playing with it.
If you are not AI-first, you are late, but you are not too late. It's a great timing to take action. All others did it, MYSELF INCLUDED, my auditors & team included. Start right now🫡
Google's proposed workaround for installing unverified apps requires 9 steps, a 24-hour wait, and runs through Play Services, which Google can modify at any time. It hasn't shipped in any beta. keepandroidopen.org #KeepAndroidOpen
Now, we have completed abstract algebra, elliptic curve. Nicolas Huber helped a lot during first half, especially in ECC.
Now, we are going to get deep into SNARKs. Most sessions will be lead by Ujjwal(@mehtasec) .
If you want to catch up, join us on the @__zkhack__ discord.
Ujjwal retweeted
MoonMath Manual Study Group incoming - kicks off this thursday March 5 on our Discord server! Find link here->zkhack.dev/
If you are a beginner & have wanted to build up your knowledge of the foundations underpining zk and SNARKs, this may be ideal for you!
completely disagree.
You should build something which actually solves a problem, doesn't matter if you take help from AI or not.
And don't be like "okay, we can build anything with AI, lets build any random stuff" , that product will be of mid quality.