grawl

Nothing more than : { border-radius: 10% }

I guess they reached the point where Best Part is No Part and Move Fast and Break Things cross their paths !

I put a dollar on a pixel-art like live projection of the descent. ** loud space boom boom plays in background **

I like learning some new code, and won't take any chances in trying to understand what jibber-jabber ChatGPT just got out. Would you please, @Gargamel, explain a bit the purpose of the Desmos language and the constructions used in the example ?

@snkiz Nonetheless, that's the way I chose to design my response. No harm or misinformation intended however, I just felt it was looking better like this, in a more concise and straight-to-the-point manner. And maybe I should have added a few smileys to highlight the tongue-in-cheek intention. Also, I would'nt call "the medical field", a narrow use case. Anyway, I'm curious about your idea of an AI learning to walk before to jump or whatever. How would you do it for, let's say, GPT-3 ? Or Watson (I'm really not knowledgeable on this one though) ?

Well of course ! They are AIs. What did you expect from them ? Maybe you've been carried away by the way they are marketed to investors and the general public.

The very issue of these AI systems is the way they are marketed. In no way they are like a person to chat with. The scope of their "knowledge" and use cases is just too broad. As @sevenperforce wrote, merely spitting words one after another, evaluating the probability of the combination and rewording if a better one is found later is not "making sense". It "making sense" is just a happy side-effect of the process. But, if properly trained, with a robust corpus of documents and a narrowed scope in mind, it can be very efficient. You can actually make your own GPT-3 talking head an expert in some specific field, by embedding context documents to the flow. As tutored here : OpenAI Cookbook : Question answering using embeddings

Overall, the only solid thing we know about is that we don't know what is aboard, as it's calssified. We can only speculate. As usual, some speculate wildly, others speculate in a more conservative fashion. The synthesis of these speculations is probably closer to the truth. Anyway, we'll know about it in, what, 50 years ? I don't remember the time needed for de-classifying stuff in the US. Time to take our pills, to last that long. Actually let's hope we won't discover it sooner, as it would mean some of the secret stuff has been used in a conflict.

Did I stutter ? ... ... Sry, I always dreamed to say this one More seriously. Yes, I understand his statements, and I do fail to understand why would they be considered absurd. As I wrote, army doing army stuff. That's their job after all. I don't think it's its sole property though. Could have been used once to test nuclear materials, another to try out some lasers, another for new fuel cell type, or whatever classified stuff you might think of. Or even all together for the same trip. Maybe just to test the craft itself as well. I'd be more doubtful about the orbital bombarment system way. But if the Pentagon decides to build a new, bigger spacecraft, it will have a lot in common with the X37, as much as will benefit from its R&D. And if they want it multi-role, it will be able to use an array of conventional and non-conventional weapons. @kerbiloid prediction might not be far off the bat in this case.

So, we do agree ! I missed your post a few pages back. Army spacecraft is doing army stuff. Not quite the surprise. I don't get the numerous posts over this question. What's the issue with this interpretation ?

@kerbiloid Regarding the long flights, I'd guess they are making long-term tests of exposure for some materials. What materials you'd say ? Probably some of them are fissile. But they can be many other things as well. Mystery goo comes to mind... Maybe some satellite imagery gear as well, but I doubt it. Maybe only for testing purposes. To answer your second question (why use the X37 ?), I'd say just because, it fits the requirements needed for the experiments aboard. And it happens to have the benefit of giving the USSF more knowledge in reusable winged spacecrafts. It's a win-win situation. Take it or leave it that's my uninformed opinion ! Thoughts ?

Is that white spot the asteroid ? 1h to impact ! Dang, I hope it will be more than 2 frames with a good view of the asteroid.

How sharp would be a look from Webb at the closer stars ? I mean, those obviously very far stars are a few pixels wide... How large would be, say, like, Proxima Centauri ?

Is that gravitation lenses we are looking at ?

For the love of god I thought that interpret would never start to move. .... Blah blah blah show us the pic already .... Show, then speak later! It's not a darn TV show. Ho, wait... ... Nice ! That's a lot of galaxies.

I remember the plan was to fully set it up in 6 months, once up there. So, I guess we may see pictures during the summer.

It just occurred to me. The title of this thread might have a totally different meaning in the future, if a more significant impact ever happens.

That makes it 15 years. Don't forget it's in Elon Time units. For the real average time in years you have to add the two values.

I'll blame it on Santa's sleigh if anything happens .

That would be one... interesting... ride !

Can a "spin" launch be detected by the traditional rocket launch alarm systems ? The tip of the rocket heating after tube exit might be enough... But if not I can see a lot of military interest in this... thing.

What's with the anime-like sword ?

To clarify as well, my answer was an attempt at a lighthearted comment with a touch of humor. In no way I believe it is a good idea to do QA on Starship this way. Final QA that is. But for a tile's laying process validation test ? Yes, of course. I thought that was obvious. Darn, maybe I should have add some smileys.

It's probably meant to identify the poorly attached tiles, and fix them. Those that held ok will probably keep holding, so no worries about them, those that got knocked off are replaced and thoroughly checked. It's not a bug, it's a feature, actually. I see it as a way faster than full inspection !

Whatever the case, moar boosters is the solution.