I often have 10+ running in parallel. I’m attacking parallel problems that aren’t interdependent. Sometimes adding additional products can bring me up to 15+.
Gotta have really good test harnesses so they can largely fix themselves.
We have our doubts about this. Can you share your code or product?
Anecdotally, my mistakes and lack of understanding exponentiate the more I try to parallelize.
As I said in the neighboring comment, for vibe coding side projects and prototypes for work I just merge and iterate. It works out more than it doesn’t. For anything bigger at work I cannot share as I’m at Apple.
But you have to keep it in your head, and remember all stuff at the same time. How is it possible to track, and do reviews one after another? Or are these pretty long running agents?
I’m not sure what you mean by keep it in your head? I know all of the parts the agents are working on. It’ll often be a mix between bigger tasks (some large refactor, new feature, etc) and small tasks (little bug fixes).
For prototyping I just merge. I don’t bother to review the code. For anything more important than I am reviewing the code and going back and forth. Basically there’s a queue of stuff demanding my attention, and I just serially go through them.
What’s also been really helpful to me is /simplify and similar code review skills (I have my own). That alone takes an agent a while to parse through everything it’s done and self reviews. It catches quite a lot itself this way.
>I’m not sure what you mean by keep it in your head?
If the project I work on is large enough, it takes me some time to get everything I need to understand for review into the short term memory. If it's small enough, it's less of a problem for me.
I don’t understand why welfare is the answer. To me it seems we’ve super failed if that’s the case — just brings everyone down except a few ultra rich people.
UBI is not welfare. It is just a livable minimum wage, for everyone who works. For those who cannot work, it replaces welfare, but that is not it's primary purpose.
As a welfare replacement, it is much more efficient, since there is no effort spent determining who qualifies. People can spent their money however they want, rather than the patchwork of separate programs we have now.
It doesn't need to bring anyone down. It's just a different way of distributing what we already receive. For you ordinary workers, they will receive $X in a monthly check, and their salary can be reduced by $X (since the minimum wage can also be abolished).
That does mean that the desirability of some jobs will shift. Good. We have a bunch of very dirty jobs being done for minimum wage, even though demand is extremely high. I'd love to see the garbage men and chicken processors get more money for their dangerous work.
And if I get less for my cushy desk job, oh well. Especially since we seem to be putting all of the effort into replacing me, and none into the jobs that come with hazards to life and limb.
The annual minimum wage (at the federal level, not counting states with higher) is around $15k. There are about 267 million adults in the US.
That is double current federal and state welfare spending.
I'm dead tired right now so I'm sure I'm missing something, but considering that is far below the poverty threshold in any big city, I dont think we'll be solving anything by eliminating welfare in favor of UBI.
UBI is basically of no benefit to the upper middle class or wealthy, and it won't be enough for the poor who cannot work enough. It really only benefits the upper lower class and lower middle class the most.
That sounds right. But I think that's a reasonable goal.
It doesn't benefit the wealthy at all. They come off worse for it. (There are revenue-neutral versions but I don't think they suffice.) But I believe that they can afford it, and will find the result a healthier America that they won't want to abandon.
But surely you can see that if the main selling point of UBI is
"Everyone gets a livable minimum wage! Oh by the way if you had a cushy desk job, that's gone because Claude can do it, or you get paid peanuts to manage Claude instances if you're lucky. Don't worry though, you can still make big bucks by working as a garbage man or at a chicken processing plant"
A UBI is basically impossible to implement on a large scale without there being significant downsides. In what world does increasing the budget by a trillion dollars or more work out well?
If the promises of AGI pan out, there will be nothing a human will be able to do better than an AI. If humans can't contribute economically, what else could things look like?
I'm guessing they reimplemented the toolbox at the TRAP level (most MacOS calls at the time were accessed through the 68K TRAP instruction).
So, rather than emulating hardware to run native ROMs, they "simply" reimplemented the ROMs.
A friend of mine did this at another level. He basically rewrote the bulk of the toolbox as a C library so that the company, who had a Mac application, could port it to run on a PC, while sharing the source code.
This was before Windows, and it worked! Launched it from DOS, takes over the entire screen. He didn't copy the Mac look and feel. Instead he used OpenLook for his gadgets and what not (since it was, you know, "open").
But he rewrote the bulk of it: QuickDraw, Event Manager, Memory Manager, Window Manager, etc. Just ate it like an elephant. I don't think his regions were as clever as the Mac. Pretty sure he just stuck with rectangles.
Correction: 68K Mac OS calls were A-line traps — in other words, they had opcodes of the form `$Axxx`. To the processor, they're unimplemented instructions that each take an exception through the same vector. The exception handler is the Mac OS trap dispatcher.
`TRAP` is a different instruction, with opcodes `$4E4x`. Each one gets its own exception vector.
It's not just trap calls, though — sometimes applications write directly to the sound buffer or use hardware page flipping.
As I recall MacOS system calls were done through invalid instructions which would cause the CPU to "trap" (raise an interrupt). Giving rise to the question Mac extension writers asked of each other: "How many traps did you patch?"
You want to talk trap patches? Your Mac's INITs have minimal trap patches, touch the Toolbox in boring places like SysBeep() and GetNextEvent(). The Developator's Mac has big, tricky INITs with trap patches that hook into half of the Device Manager.[1] The Developator is in touch with the metal, starts like a warm boot off the ROM reset vector, stops on an NMI.
The question serious patch trappers ask is whether you patched the traps used to patch traps to make sure that, when you patched traps, no other INIT could patch those traps after you to get in line before you when the traps were handled.
Hey, I’ve been thinking about building a genesis emulator to play Revenge of Shinobi. Never thought about trying to replace the genesis rom.. let’s get the game working first :)
I cannot tell you the number of times I’ve been in France and had people correct my pronunciation (which btw, it’s really not so bad — the best complement I got was that they could tell I was foreign, but not sure where from).
One should never draw a redline they aren't willing to cross. Trump of all people should know this, he gave Obama shit for years over the uninforced redline with Syria over chemical weapon use.
To Trump, when someone else does something, it's worthy of reproach, but when Trump himself does it, it's the cleverest 4D chess anyone could ever imagine.
A more fundamental aspect of his character is that everything his enemies does is bad and stupid and everything he does is good and genius.
So if Obama allows red lines to be crossed, it's totally different. Trump has a long history of bluster and hyperbole and - look - we re-elected him so I guess it can be a winning strategy.
Or it was. Now that people are calling his bluff on tariffs and genocide, there isn't so much winning.
reply