Chinmay

[Chinmay @ChinmayIO]

@hackaday Ain't nothing gentle about it. Its pure wild ..fun :D

[Chinmay @ChinmayIO]

@HarryStebbings

[Chinmay @ChinmayIO]

@lennysan I completely agree with that comment.

[Chinmay @ChinmayIO]

@pitdesi lol!! I think similar story as Madras textiles for bleeding dye. At least they gave credit to India unlike the "Scandinavian Scarf"

[Chinmay @ChinmayIO]

@MatthewFalls6 @osquitardotcom @SahilPunamia @dieworkwear https://t.co/RanevxYzeo

[Chinmay @ChinmayIO]

@ShanuMathew93 So I am assuming a higher % for computing hardware..guessing 70% now. Energy costs for building a cluster are small (<10%). Of course, when you run those clusters, it is completely different

[Chinmay @ChinmayIO]

@ShanuMathew93 For Capex, 60% compute hardware, 25% cooling, electrical, and networking. But this was for DCs for cloud operations when I worked in data center efficiency. Since the power density of GPUs is 2x to 3x CPUs, more cooling equipment will be required, but it won't scale linearly.

[Chinmay @ChinmayIO]

@mervenoyann Thanks for releasing it. I just tried it, just to confirm, it only has English language support, correct? I tried it with a Telugu and English mix document, and it said it was not trained for it.

[Chinmay @ChinmayIO]

@andreasklinger You are sane/insane enough to start a company so they are just asking you one more time are you sure masochism is the path you want to choose. The whole activity is the induction to what's to come on this path.. which is eating glass everyday, :D

[Chinmay @ChinmayIO]

So

[Chinmay @ChinmayIO]

I did not expect this kind of populism and regionalism from Mahua Moitra. If these allegations are true, the people who took bribes are equally responsible. I do not see her calling them out by their states.

[Mahua Moitra @MahuaMoitra]

Modi and his Gujju cronies have made India Inc. look like a bunch of chors and thugs. Terrible for our credibility in global markets. Can any large business house now have the guts to speak up?

[Chinmay @ChinmayIO]

This also slight their work and contribution.
Hopefully one day we will be able to go beyond this regionalism, but clearly today is not the day.

[Jordan Taylor @Jordan_W_Taylor]

Moonshots & mass production: The lesson of Reaction Engines.

The loss of Reaction Engines, the British architect of the incredible SABRE engine, has drawn unfair comparison with Musk’s Starship: Some say it renders SABRE obsolete. Is this true, or idle midwittery?

It's true, but not for the reasons you think.

Comparisons between SpaceX and Reaction Engines miss the point, because they're very different organisations. One is an industrially scaled launch services company and the other was an R&D outfit focused on an all-or-nothing moonshot: A powerplant for a perfect orbital delivery system.��

And the perfect is the enemy of the good.

We'll say at first that SABRE is a spectacular concept: An air breathing hybrid rocket engine combining performance with incredible fuel efficiency, it would be a game-changer if married to a suitable spaceplane that, admittedly, doesn't exist yet, to supply a high intensity orbital market that also doesn't yet exist.

Difficult!

It would use cryogenic hydrogen fuel as the heat sink for a helium chiller loop using the world's most extreme heat exchanger as a super-precooler to let an air breathing, turbo-compressed engine work at up to Mach 5, then switch seamlessly to rocket mode.

Got that? Good, because that's just a fraction of the challenge behind the SABRE engine, or Synergetic Air-Breathing Rocket Engine, to use its full moniker. No wonder Reaction Engines never finished it.

That's unfair. They had funding in the millions for a project that needed billions, so they did what they could: They focused on developing the most challenging subsystems; helium loop, heat exchanger, rocket motor & preburner. The full solution was beyond them.

Even then, however, there were gleams of pure brilliance: With their near-miraculous precooler they charmed a modified Rolls-Royce engine into running in Mach 3.5 conditions, something even the SR71 Blackbird's unique powerplant couldn't quite manage.

SpaceX, by contrast, approached its challenges as the industrial business it is. We forget it now, but its first focus wasn't giant reusable spacecraft, but cost engineering and design-for-manufacture: It brought key production steps in-house and relentlessly optimised them to drive down cost. Even the Merlin engine is a case in point: A small standardised engine, many to a booster, that could be produced at rate. It meant that, with more engines per rocket, SpaceX would benefit from industrial learner effects quickly and gain reliability and failure rate data faster than its competitors too.

So cost engineering drove manufacturing repeatability. Repeatability then drove increased reliability, and only then was increased reusability a realistic goal. Crucially, each step gave it another advantage over competitors, drove down launch costs, raised profits and increased cash flow, financing yet more improvements.

And it's why SpaceX has hundreds of times more resources to throw at difficult problems than Reaction Engines ever did. A big anchor customer in the form of NASA was useful too.

We fetishize blue-sky innovation at our peril. It's important, but far more important is the ability to scale; to bring your innovative genius to customers, make use of it and spin-up the kind of industrial flywheel that pays for yet more innovation down the road. This is how industrial capitalism should work, and it's something America is very good at. 

SpaceX is a uniquely American company in that way. Who knows if Starship will scale-up like the rest of the business, but there's a logic to it. You can hear the flywheel turning.

Reaction Engines, then, might be a uniquely British company: Incredible brilliance deployed at a problem in the future, for a market that doesn't exist yet.

A moonshot that won't reach the moon.

Sometimes you don't want mankind to take a giant leap: Sometimes, a thousand steps will get you there faster.

Blue Origin got its motto right: Gradatim ferociter.

Step-by-step, ferociously.

[Chinmay @ChinmayIO]

Proud to support this petition for a unified pan-European startup entity. Let’s unite Europe's startup ecosystems! https://t.co/e1y3S2AFn9 🇪🇺 https://t.co/x9rvknwQCo

[Chinmay @ChinmayIO]

@tobi https://t.co/upwJ9c3m05

[Chinmay @ChinmayIO]

@sriramk https://t.co/upwJ9c3m05

[Chinmay @ChinmayIO]

This is not as cool as he thinks it is.

[Thierry Breton @ThierryBreton]

Deal!

#AIAct

[Chinmay @ChinmayIO]

@AnthropicAI 's Claude is so neutral that it recommends the @OpenAI GPT4 model for summarization.

[Chinmay @ChinmayIO]

@mouthofmorrison Jira literally means Cumin in several Indian languages, so double whammy. I could literally do a sprint on these kind of names. What an epic time we live in when these stories highlight different linguistic issue types. Though, I would see it as a feature not a bug.