gradient descent..it's not just an algorithm..it's a lifestyle

Thank you for the good, original content

wow Nikita, thank you so much 🤍

@nikitabier Good job so far 👏

@nikitabier thanks Imran!

this way of seeing life - and the way you explain it - is so, so resonant for me. much more fresh and interesting than the usual fortune cookie style life advice one sees on social media.

ty Patricia that's great to hear

We need more content like this. Thank you!

thank you!

I love this both as a CS person and as an introspective person

:)

Can you do a video on the expectation - maximization algorithm now? This got my juices flowing 🔥🔥🔥

👀

Crushing it!! Love the drawings.

thanks Ray!

Great explainer, more please @ashebytes this was valuable, thank you 🙏🤯😎

thanks Tom!! I appreciate the encouragement.. I am not sure how many reels to do on X

I think more is best. You're great. Do more. Gradient descent it!!!!

:)

it’s the old analogy- you may strike the rock a hundred times without seeing progress, but the hundred and first hit succeeds because of the first hundred 🤍 it’s the compounding of persistence that pays off

optimize for feedback before depth/direction is what im picking up

maximize for learnings to update your "weights"! can even be your own reaction to what feels good to put out vs external reception importantly, in the beginning of your optimization process as you start out on approaching the goal state.. take big bold steps in many directions. narrow and refine after you've gleaned the learnings

I am impressed how many good content creators use headphones for convenience, somebody should invent a plugin clip to make it prettier and easier to hold it

the audio for quick app upload is actually more normalized than dji

I guess it’s a perspective to perceive gradient descent as looking in all directions. Gradient descent find the steepest slope and goes in that direction and more often that is the closest minimum to the starting point. Contrary to life, you can start gradient descent at random points and get to different minimums but it’s difficult to start life at a random point simultaneously.

this is why I say "optimization process" or "goal you're looking to achieve"

This was actually really useful 🙏

:)

Followed immediately

this is so cool!

ty himanshu!

I would prefer to take the path of the stochastic gradient descent sometimes you need to try radically new things or you can be in the risk of getting stuck in a local.

what is your tech stack for editing videos? love them!

LOVE IT

REAL ! Love this !

@SynbioUnsiloed Saw this and thought of you.

this is nice content on my timeline.

I've always said "doing it for the lols" but doing it for the plot sounds so much cooler lol

if you listen closely you can hear women everywhere in manhattan whispering to each other "do it for the plot"

New York is probably the greatest place on earth to be plot maxxing

Love the content, there is a lot of elons in this, how he takes larges steps. I tend to do this since was I child, or at least think about it, but encouteres tremendous resistence and 99% failure. Really ejoyed and hopefully this video hits to all ppl here! 🙏🙏

thank you!

and searching for the right mental equilibrium for stressless creativity

I have no idea what gradient descent is lol. Never heard of that term before.

now ya know!

I have been enlightened lol.

Very refreshing. So inspiring. Thank you from Germany👋❤️

ty Julian!

❤️

Nice gradient decesent and what about leakyRelu or Relu?? Could you make video on this topics.

interesting

Might have to write a song about this, how’s that for relational

this is amazing, ashe! i love it

thank you! :)

New favorite follow on X. Great content!

wow ty Prescott! much appreciated

love it

This fits with what we have learned about psychology and success. Keep trying new things. Learn from the things that don’t work but don’t get discouraged by the things that didn’t work. Those are lessons, not failures. Failure only happens when you give up.

when a wave peaks, it descends, and on its descent, it catches a reflection of itself just long enough to be genuinely surprised its you.

Great advice! thank you for sharing!

ty for watching!

Nice simple straightforward advice. I like it.

In the beginning you take small steps to avoid instabilities caused by mismatch between initial conditions and the distribution.