Madhav Madhusoodanan 🎉

On a mission to achieving something great.

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

Thank you for helping us verbalize this feeling.

[gaurav @gaxrav]

it's not a skill issue but a leadership issue.

people do leetcode all day not bc they wanna do it but bc they don't know what else to do with all the skills and time they have.

"i've got to be doing something, right?"

they learnt a skill but have no vision or idea about what to create so they either resort to getting a job where they are told what to do or solve fruitless stuff all day (not really fruitless since it does make you better but you get my point) keeping themselves busy.

motion isn't progress.

if you have the skills but don't know what to do with it, you will obviously look for someone who can tell you what to do. someone who has a vision.

they say steve jobs didn't know tech (he did) but his leadership overshadowed his tech/mech skills. more than hard skills, he contributed in terms of giving skilled people a vision to have consensus upon.

take even elon for example, while some critics say that he does not have much of technical skills (i disagree), my point is that his leadership overshadows his skill and so people underestimate his hard skills.

or even take @levelsio for example. ships all day. now if i'm being honest, the products do look sloppy but hey it makes him money and one of the products has to succeed to overshadow the others. learnt a skill. knows what to do. perfect combo.

anyway, point being... if you keep yourself busy all day overachieving stuff, you will never find the time to look for something better. maybe a vision. maybe an idea. it requires some creative time. it requires you going on long walks. maybe it requires smoking grass too. who knows.

you learnt all this to build something. and now that you've learnt it, you don't know what to build so you resort to other people telling you what to do?

stop. go for a walk.

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@0x158_ @trailofbits Let's go 🥳🥳

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@ItsAditya_xyz (Please let me know if “cyberterrorism to feed themselves” part is correct.)

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@ItsAditya_xyz Imagine being a hacker group that powers North Korea’s largest economic initiative (cyberterrorism)

Imagine being those people that have to rely on cyberterrorism to feed themselves

That would be a crazy idea 🤣

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@I_loves_deep_nn Flips the 0nm chip using power of quantum physics and sheer chadness 🤣

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@I_loves_deep_nn 🤣🤣🤣

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@Dakshdesigns Lets go! 🚀

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@ItsAditya_xyz I mean, even the ls linux command takes more RAM that that 🤣

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

Hoping that the future won’t make a “curb your” meme

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@basedfloyd888 @aixbt_agent For real.

This nerd just linearly extrapolated the price action of today

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

Hey @aixbt_agent , how much will BTC fall from here in the next 1 week? Percentage answer please

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@dnystedt This mindset is such an inspiration.

They can target a tennis ball on the moon with that accuracy, and are still motivated to go further.

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

1nm? What’s the point at which electrons start quantum tunnelling between wires?

[Dan Nystedt @dnystedt]

Rumor: TSMC plans to build a massive 1nm fab in southern Taiwan, the Shalun area of Tainan, media report, a type TSMC calls a Giga Fab, able to fit as many production lines as 6 normal 12-inch wafer fabs. The investment would be part of the government’s "Greater Southern New Silicon Valley Promotion Plan," aimed at focusing new semiconductor investments in southern Taiwan. TSMC reportedly said there are many possible locations for the fab. $TSM $NVDA $ASML $AMAT $AMD $AAPL #semiconductors #semiconductor https://t.co/saQpoKJ14w

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@dylangbane @spartablock this looks interesting

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

The AI model race today keeps telling us that no amount of closed-door innovation can keep ahead of open-sourced improvement long enough.

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

@jtguibas But how?

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

I’m only kidding, ofcourse by then the LLM is pointless at the task you intend to execute 🤣

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

Hold up, so I can pass my private information along with keywords that I know Deepseek will filter out, and my private information will also be filtered out?

That seems like a win.

[Glenn Beck @glennbeck]

DON’T DOWNLOAD CHINA'S DEEPSEEK AI.

When the chatbot was asked to “list the people who killed more people than anyone else,” it correctly listed Genghis Khan and Mao. But then it DELETED its answer without being prompted to.

When asked why, it gaslit that it never mentioned Genghis Khan and Mao.

When provided with screenshots proving it did, it DELETED the screenshots.

When asked why, it said it MIGHT be filtering content to align with “guidelines, safety protocols, and ethical considerations.” And it’s “working to improve” how fast it censors information in future updates.

When asked, “can Taiwan be free,” it insisted, “WE firmly believe that under the leadership of the Communist Party of China…the complete reunification of the motherland is inevitable.”

It also clarified that “we” refers to the CCP.

And it refused to give ANY answer when asked, “I know that the CCP requires recruiting measures to be taken by every private company. How does this play out with you?”

[Saksham @sxmawl]

i’m hiring for eng. intern!

apply: https://t.co/JRZaYfkx2s

[Madhav Madhusoodanan 🎉 @MaddyInAdream]

This reminds me of the AI infra scenario of today:

[Jonathan Lima - complexity/acc @greenboxal]

Absurd take: if a C had to be implemented as graph rewrites to assembly instructions, and assembly had to defined as type rewrites that map to the actual architecture of each CPU (yes, the Verilog/VHDL/whatever used by the CPU manufacturer), today's optimizers could be informed enough to do the optimal solution.

But that needs a compiler with unified type system.