You've misunderstood. The parent was making a specific point — if you want an SVG of a penguin, the easiest way to AI-generate it is to get an image generator to create a (vector-styled) bitmap, then auto-vectorize it to SVG. But the point of this benchmark is that it's asking models to create an SVG the hard way, by writing its code directly.
When I first heard all this about honey I was shocked, remembering seeing Linus plug them. Of all the people to have the potential ability to see through it. The way I see it is that anyone who sponsors things like YouTube videos as widely as they do is generally a piece of s** company. Normally up to something, that makes it worth their while to spaff money on such things. 80 quid razors, AI driven news classifiers, VPNs, meh...
My more general rule is that anything being advertised to me must be way overpriced or a scam, in order to pay for the expensive advertisements. I won’t buy nearly anything I see advertised. I don’t run into many ads anyway, but some always get through!
Well he brought down his entire storage system. Twice.
I believe one time he had to bring in Wendell Wilson from Wendell Wilson Consulting, but more likely know to the Internet-at-large as the primary figure on Level1Techs YouTube channel.
I have no desire, nor inclination to dig through the thousands of videos LMG has produced on YouTube, but it's still up to my knowledge, and watching him fumble about with absolutely no clue is not only painful, but pathetic. Linus suffers from the same affliction many of my Ph.D.-holding friends have, which is that he believes because he knows a lot about putting together computers and electronics that he could handle building a large-scale data storage system.
These systems are complex, and to be well-built and maintained, they require domain-specific knowledge - no different than an OS programmer needs deep knowledge of C and C++, and increasing now, Rust.
It's a series of videos of someone way in over his head who should have brought in an actual expert - like Wendell from Level1Techs, or Patrick from ServeTheHome, from the get-go, instead of trying to do it himself.
Here I have to chime in and say that a certain YouTube razor is one of my favourite purchases ever. But I guess it's rather niche, being a double edged safety razor.
Come on, this is basically what a Dualit Classic is. I really don't see this is much more than an exercise. The problem is asking people to spend 5x to 10x the price of cheap mass produced unmaintainable tat.
I believe Amazon impose a whole bunch of conditions on people publishing on their platform. Along the lines of exclusivity and controlling where else if at all you can sell it. I'm pretty sure Corey Doctorow extensively covered this.
Yes they do. One particularly annoying one is about pricing: price can't be zero. If you write as a hobby and don't care about making a dollar, then Amazon gets in the way. Sure, you can publish in Amazon and charge 1 USD, but then you can't publish your book on your own website for free because it goes against Amazon's TOS.
There are other peeves. Covers for example: it's against the TOS to have a cover that shows female nipples, but it's okay to show male nipples. Beyond the sexism of the rule, I'm worried that the way to enforce this is to have some ML system checking all the covers and making judgements about nipples. Which means you have to ask your cover artist to not draw anything that may accidentally look like the wrong kind of nipple \o/ .
First attempt to catch the booster back at the launch site.
The "mechazilla" launch tower has two "chopstick" arms which are used to pick up and stack both stages and which are intended to be able to catch the returning booster and maybe also the returning Starship upper stage.
Yes. The booster has two pins that stick out at the top that are designed to hold the weight of the entire booster when empty. The plan is for the booster to return to the launch tower, position itself between the arms which will close on it and then the pins will “land” on the arms, completing the catch.
I’d say the main difference, then, is that the booster will be supported by those pins resting on top of the arms. Chopsticks use friction to hold up their load.
yes the booster’s structure is very strong vertically but not nearly as strong horizontally. There may be some “squeezing” forces from the chopsticks but this is effectively for fine positioning only. It will not support the weight. The booster will “land” by getting its pins (which stick out a bit) on the top rail of the arms.
It should be better described as having the booster land on the arms. The arms will probably be able to adjust a little to assist in alignment, but the booster is doing most of the work to be 'caught'.
How could it possibly be meant literally? Do you consider it possible for a rocket to be caught by a literal person with literal wooden sticks?
I guess I don't really understand what you are asking. There's a tower with some huge metal arms that is meant to catch the rocket. They call them chopsticks in a joking manner. Obviously, I would have thought.
Yeah I totally envisioned a person holding wooden chopsticks trying to catch a booster /s
You missed the quoted part about > which are intended to be able to catch
Which would be the unique thing to clarify. As in "something like" the "chopsticks" moving to > catch < the thing -- Like Mr. Miyagi moving the chopsticks to > catch < the thing
It allows removing the landing gears on the booster, which saves wheight, which saves fuel, which increases efficiency and reduces costs. It also avoid having to fetch the booster from wherever it would have landed.
Given that a lot of the landing failures we've seen started with a near perfect landing followed by the rocket tipping over, I suspect one benefit is that the contact point is now above the center of gravity and thus it can't really tip over.
Of course, it can't tip over unless something fails or the rocket ends up in the wrong spot (and fails to get caught) and the previous tip-overs also had to involve failures (of the landing strut, in the latest loss) or landing in some way that isn't perfectly aligned.
Their launch license requires them to initially aim at the water, and only shift to aiming at their tower if both the booster internally judges it's in perfect health, and they send the signal from their control system.
I think there is a reasonable possibility that something goes wrong enough at some point for the booster to go in the drink. But if that happens, maybe it'll be close enough to the shore that we'll get some nice video of it?
This is also standard procedure for Falcon 9 landings. They would do it this way even if the launch license didn't require it, because they know the probability of some sort of failure of the booster is high, and they don't want to destroy the launch tower if they can help it.
"According to the National Fire Prevention Agency, if an EV ever catches fire while you’re behind the wheel, immediately find a safe way to pull over and get the car away from the main road. Then, turn off the engine and make sure everyone leaves the vehicle immediately."
They should have written turn the ignition off ("ignition" also being a misnomer for EVs!). That will disconnect the battery, which won't stop the fire but will make firefighting slightly safer.
