Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror

Comment Re: No [to whsat question?] (Score 1) 107

But back to the original story... I am inclined to agree with him, which is why I'm trying to limit my use of AI tools and also trying to consider how such tools affect my approach to solving problems.

I don't think you need to worry too much, you have a sufficiently analytical approach that it's not likely to be a problem.

Comment Re:Fair Comparison (Score 1) 23

He probably is. I would too if I had his ego and told everyone it'll cost less than $60 million. But at the end of the day, it's just speculation.

I think the true cost is not too far off, i.e. within 3x of the advertised price. They've launched ~90 rockets, so $20 billion in government contracts is ~$200 million per rocket launched. Normally, businesses spend about 20% in management, and for high tech industries, 20% in R&D. That leaves about $120 million for each rocket. Of course they also made some profit from private launches, but that's offset by the amount of cash they have on hand and a number of rockets that haven't been finished

Comment Re:Umm... (Score 1) 60

2. These examples 'play' chess in the form of a textual representation of a sequence of moves, not 2D representations of the board. That means that the AI needs to 'mentally' replay the moves to determine the state of the board. It's like asking it to do 20 consecutive math operations on a matrix of 8x8 and then asking it to come up with an operation that leads to a specific type of state of the matrix.

What you are describing is a good way to play blindfolded chess. Every time someone makes a move, repeat in your mind all previous moves from the beginning until the present. I've taught several people to play blindfolded chess this way (and also learned it myself, from George Koltanowski).

Comment Fair Comparison (Score 4, Insightful) 23

The Qianfan program has existed only since Nov 2023. So for a fair comparison, we should look at the first 1.5 years of Starlink. Starlink launched ~960 satellites between May 2019 and Dec 2020. The satellites were 250 kg each. Each Falcon 9 was able to launch 60 of them, so 16 launches. Meanwhile each Qianfan is about 300 kg, same weight as modern Starlinks. The Long March 6A that they use can only carry 18 of them. They have 90 satellites, so it's just 5 launches.

Around 2019, most Falcon 9 boosters could only be reused 2-3 times, with a cost of $50-60 million per launch taking reusability into account. There's no information on the Long March 6A, but the Long March 6C costs around $70 million. There simply isn't a 3x difference in per-launch cost that could explain why Qianfan only did 5 launches as opposed to Starlink's 16.

My guess is they're waiting on the development of cheaper launch systems before spending a lot of money on it. The Long March 6A is made by the government and costs a lot more than it should. There's several private Chinese launch companies that are developing low cost reusable rockets. At least two have completed soft landings, one of those companies has a rocket that reached orbit (though with a disposable one). Check back in a year or two and I suspect the numbers will be very different.

Also keep in mind SpaceX started the reusability program in 2011. It was 2017 when reusing the rockets became economically feasible (i.e. costs less than a new rocket to refurbish). Looking at the advertised price of a Falcon 9 today, I think reusability might have cut the cost by 20% or so at most so that's not the main reason there's been so few Qianfan launches.

Comment Re:This is not an AI failure (Score 1) 148

Still has the word "elephant".

Rather than countering GP's argument, you actually proved it. Adding the instruction actually causes the model to veer towards patterns that relate to the instruction itself, because it's lacking the ability to understand the words you're using to negate the rest of the sentence.

Comment Re:Prurient Interests (Score 1) 136

As the meme says, "we knew more about the Coldplay Kisscam couple after one day than we know about Matthew Crooks after one year."

There's a lot of info about him.

Would be more accurate to say, "We know why the Coldplay couple did what they did, we have no clue why Matthew did what he did."

Slashdot Top Deals

On a clear disk you can seek forever. -- P. Denning

Working...