Khris Rino

engineer • photographer • travel enthusiast • tech geek

[Khris Rino @khrisrino]

Jupiter doesn't orbit the Sun https://t.co/BK2t3cROlR via @YouTube

[François Chollet @fchollet]

"Any software engineer can now produce much more value than before"

"No one will want to hire software engineers"

Both of these cannot be true at the same time.

[Khris Rino @khrisrino]

@levie Sometimes founders get all the credit ignoring the blood and sweat of the lower layers. It's easy to reduce everything to LOC (lines of code) or some other convenient metric but it's usually not the entire story.

[Lukasz Olejnik @lukOlejnik]

Chinese scientists have developed the best shortest-path algorithm in 41 years! A team from Tsinghua University has broken Dijkstra’s “sorting barrier” — the first improvement since 1984.
The new algorithm runs in O(m \log^{2/3} n) time. Potential applications? Faster shorter waits for route calculations, fewer traffic jams, cheaper deliveries, and more efficient computer networks. And, of course, a need to update computer science curricula :)

[Khris Rino @khrisrino]

@vasuman One shot refactoring is usually a disaster … even with current gen LLMs. For complex code base I prefer iterative bottoms up approach. Carve out small chunks of code at a time

[Earth @earthcurated]

Alaska ❄️

[DHH @dhh]

The creative process can't be planned, and it has to be difficult. Otherwise the answer would be obvious. The key insights often don't arrive until you're halfway through the implementation. Wishing you were smarter/wiser on day 1, so you could have avoided the twists, is silly.

[Paul Graham @paulg]

AI startups seem to grow faster than previous startups. But this is just the latest stage in a consistent trend over the last 50 years. Each decade's startups grow faster than the previous decade's.

[Balázs Kégl @balazskegl]

This little guy is conscious. Tell me why not.

[Khris Rino @khrisrino]

@Altimor Are we talking shortest by time or distance here? One issue with AM is they optimize aggressively for best time so the route can be many more miles just to save a couple mins. If I’m not time constrained or have a preferred route I prefer GM.

[Valeriy M., PhD, MBA, CQF @predict_addict]

Kolmogorov-Arnold network obliterates Deepmind's results with much smaller networks and much more automation.

KANs also discovered new formulas for signature and discovered new relations of knot invariants in unsupervised ways. Incredible 🔥🔥🔥🔥🔥

The bell 🔔has tolled for deep learning finally?

#kan #deeplearning

[Nat Friedman @natfriedman]

Instead of leaf blowers, I want a quiet little robot that picks leaves up one at a time and puts them in a bag, at night while I'm sleeping.

[Khris Rino @khrisrino]

@LinusEkenstam The only reason for a self driving taxi to be preferable is that it’s cheaper or takes you to your destination faster. Any perceived privacy or safety benefit needs to be thought through one more time.

[Khris Rino @khrisrino]

@fchollet The main issue here is we’re trying to model an infinite number of things simultaneously. Even if it gets the physics and common things like human faces accurately enough it will still always render weird things it has not seen enough data on. Can we get it to avoid the latter?

[Khris Rino @khrisrino]

@neilcybart A product is not just a spec sheet. Sometimes the secret sauce is that bit of magic … that extra mile the engineers go to deliver a great experience. Clunky products don’t win.

[Andrew Côté @Andercot]

Under water fiber optic cables are cool but have you considered just beaming information straight through the mantle of Earth using a Neutrino Comms Laser?

[Yann LeCun @ylecun]

Too often, we think a task is easy because some animal can do it.
But the reality is that the task is fiendishly complex and the animal is much smarter than we think.
Conversely, we think tasks like playing chess, calculating an integral, or producing grammatically correct text are complex because only some humans can do them after years of training.
But it turns out these things aren't that complicated and computers can do them much better than us.

This is why the phrase "Artificial General Intelligence" to designate human-level intelligence makes absolutely no sense.

[Khris Rino @khrisrino]

@cunha_tristan @fchollet I want to train an LLM on this debate and see the outcome :)

[Massimo @Rainmaker1973]

Emperor penguins survive temperatures of -60°C thanks to their feathers, densely packed as seen in these pictures.

They have the highest contour feather density of any bird, and both filoplumes and plumules (downy feathers).

The downy plumules are four times denser than afterfeathers and play a key role in penguin survival.

[M. Alex O. Vasilescu @AlexTensor]

I agree. Prediction does not imply that an #LLM has learned a model of the underlying reality. It does not imply an understanding of language, human thought or physical reality. Understanding is causal in nature. #causality

For example, the Mayans were able to predict with great accuracy when eclipses would take place, but without an understanding of what the moon was or that it moved around the earth.

Feynman video: https://t.co/H2SiOO2Qn5