Tiny team; big technically challenging project across a several compute domains from microcontrollers to S3 buckets. We don’t really do real vibe coding except on web. (Basically static site)
App development: 4x. Lots of interface.
Firmware development c++: 10x+, with a huge increase in code quality. This requires extremely methodical planning, where planning and specification is 80 percent or more of the workflow…. But the results are outstanding. The “trick” is that planning and specification is basically code generation. By the time something is fully specified it’s basically done. The LLM just allows the humans to work in a more productive layer of abstraction and also offloads a whole lot of burdensome calculation and research.
Backend development: 10x at the prototype level it’s basically 100x magic. Everything one-shots if you go with simple python based cli tools with tests. Then you string them together to make the workflow. Once it works, you go over each tool with more testing and optimisation, library deletions where practical, etc. By building each significant function as a separate cli tool it makes testing/fuzzing automatable, simple and reliable.
It’s not that simple and we shouldn’t be so eager to blame our neighbours.
Trump has been heavily involved with Russian money ever since US banks refused to loan him more money for his repeated bankruptcies. At that time, Putin was head of the KGB, iirc (could fact check this I’m not 100 percent sure). Also very sus ties through marriage. I’m not sure he is -happily- compliant, but from whatever combination of stick and carrot that may exist, there is little doubt in my mind that the administration is carrying so much water for Russia that europa would blush.
If you want to see the whole story in embarrassingly human terms all you have to do is watch the physical interaction between Trump and Putin when they are in close proximity. The monkey doesn’t lie.
If you pay attention, almost everything the admin does benefits Russia in some way. Eroding soft power. Undermining nato. Crippling us debt. Hamstringing innovation through reckless trade shenanigans. Undermining trust. Abandonment of Ukraine, an erstwhile critical ally. Hand wavey condemnation of Russia but with no teeth whatsoever. Collapsing the straits of Hormuz. The list goes on and on.
The object is to create metrics that are plausible (or more specifically for which the knowledge of falsification is plausibly deniable), not to actually self deport people (although that could be viewed as a desirable side effect).
We can be sure that the weakness in this scheme is being exploited, since it was designed to maximise metrics, not results.
Idk, in my hometown, jail is seen by some as being preferable to winter village life. A few people commit petty crimes for the purpose of a 6 month lockup until spring (or at least this was the case 3 decades ago).
Could be that Alaska has (had) particularly great jails?
This isn’t a serious project, in the terms of something that will disrupt warfighting. It’s basically a resume to work as a junior engineer at Anduril.
Interesting stuff, neat project, nothing new at all here except his multi camera sensing, which isn’t new but his implementation is interesting.
IDK if maybe it’s a political statement or some kind of obtuse sarcasm, but it seems like he drank way too much of his buzzword cool-aid lol. It’s probably just a job application though.
UBI is the actual solution, and is well understood enough now to know that most of the arguments against it are moot points or simply falsehoods.
Unfortunately, with regulatory capture at near 100 percent and electoral capture almost as bad, there is no incentive structure with sufficient influence to make it happen. Wealth will continue to be funneled to the top, and taxation schemes that act as a de-facto sales tax create incentives that favor even more centralized systems.
But wouldn’t it be great?
An interesting aspect is that I am constantly observing innovators with significant technical and technological skills that are employed in fields outside of their expertise as a “temporary “ measure that often becomes permanent if they get further encumbered, simply because they can keel out an existence while trying to build the next cool thing. So we are wasting probably trillions of GDP in talent because people need to go work in a labor job to support their wife and child instead of continuing his very promising project in training data for humanoid robots, which could easily net 100m+ in the next decade. (Actual example. I offered him $1000 a month to keep on it, but he unfortunately needs more to survive and he has eaten through his savings over the past two years of working on it.)
Police unions aren't the same thing as other unions. Most unions exist to equalize labor negotiation through collective bargaining, and police unions tend to include and align with the leadership in the organization that a union would traditionally be negotiating with. In practice they're a lot more like a military contractor than a union (in that their role is to prevent public accountability)
No, the core premise of a union is to represent the interests of its members. You've confused a union with global communism, and delegimitized the ones that don't serve your interests instead.
For an extremely salient example, the purpose of coal miners' unions is to serve coal miners. The interests of coal miners don't necessarily align with everyone else's interests, and if the the interests of their union did, it would be a bad union.
If police are racist, police unions are either racist or not representative. The purpose of unions isn't to serve the purposes of upper middle-class liberal arts majors. You don't steer workers through their union, that's evil. If you want coal miners to prioritize the climate, or police to prioritize civil rights, you have to do it the traditional way - by convincing them. You might have to convince them to quit.
The idea that the police don't deserve a union because police unions would support your enemy is repulsive. Change the laws, maybe you wouldn't have to hire scum to enforce them.
I'm sure there are other things that call themselves unions but don't serve the same function we expect of organizations using that word. I'm not really knowledgeable about either of the things you mention, so I don't know if I'd view them as fitting this description or not. Both seem unlikely to have the kind of broad negative impact police unions have, which is particularly egregious because they effectively make public oversight of a government function with a lot of potential to do harm impossible, which has the knock-on effect of making that harm more able to propagate, to the point where it's both quite severe and commonplace
I can see you're no fan of nuance. I pointed to what I believe is an important distinction between the general function of unions and that of police unions, this is hardly a claim of unimpeachability. Then when you for some non-sequitur reason started "what about"ing other unions, I said that I don't know enough about those to know whether they have a similar problem
I get the sense that you feel strongly about this subject, but it would do you some good to read the messages you're replying to, as not doing so makes you sound pretty foolish
This reminds me of a trading incident decades ago. We were unwinding a portfolio of correlation trades. The trades basically bet on continued correlation between assets. Unwinding those trades means assets which typically correlate now correlate less. When we fucked we they sometimes anticorrelated. Anyway, some of these price movements were noticed and incessantly commented on by CNBC pundits. Theories abounded. All wrong. Most amusing.
lol. I can see the humor in that. I can’t say I have seen it IRL but it does strike me as entertaining to imagine. Sort of funny in that “haha didn’t happen to me” kind of way.
Unions just create an us vs them mentality. The fact that the NUMMI plant ( https://en.wikipedia.org/wiki/NUMMI ) was not reproducible is a pretty strong indicator of that.
Sorry, my AI autocorrected “ubi is” to “unions”. When I came back to all these “unions bad” comments I was like WTF are they on about??? lol.
But, I will say, often, unions are very positive things. Other times, basically evil. So, I think unions are a mixed bag. But the threat of potential unionization is almost a universal good, so it’s a right worth exercising once in a while.
I’m working on that, but there’s an incentive alignment problem for early stage investment capital that requires investors with a little longer view to avoid kneecapping the value proposition, as well as the potential positive aspects to society.
(Am I allowed to be a little ideological when we are facing the potential extinction of humanity?)
OK, but the same response still applies. There still are plenty of arguments against UBI, and you still need to actually refute them rather than just dismiss them with "they're false or moot".
I’d agree with you in a debate, for sure, but I’m just stating my opinion. Of course, opinions are like assholes, everybody’s got one lol, so I’m not saying im necessarily right.
But I do think in the special case of self maintaining, self manufacturing, intelligent servitors, I think the only way we can hope to sustain an economy of any kind would be to redistribute the output of those machines significantly, or we would end up with a concentration of power around capital that utterly eliminates the economy as we presently understand it.
As for moot, I mean that in this case, you aren’t taking anyone’s work to benefit others, so the usual arguments of socialism or wealth redistribution don’t have the same basis in injustice. Ultimately, the production of pure automation springs directly from the resources of the earth, and the earth is everyones, so it makes sense to redistribute the production to a significant extent.
As for false, I mean that the “no one will do anything” and other claims about UBI destroying productivity have all been refuted in study after study, and in the societies that already practice UBI. In this way, UBI is very distinct from needs based welfare, which is often imagined to incentivize low production, since low productivity is actually a requirement to qualify.
Ubi causing runaway inflation is another example of a thoroughly refuted claim that is easily given the lie by looking at extant UBI systems.
UBI seems to work best when it significantly removes or even eliminates stressors of survival (basic housing and food, medical care) while leaving lots of room to aspire to greater success. Throw away the stick, but leave the carrot.
It’s basically mirroring what behaviorists find time after time, that positive reinforcement is more effective than negative reinforcement. Also, it makes people more willing to take risks like starting a business, getting an education, or having a family. All of which are positives for developed nations.
> But I do think in the special case of self maintaining, self manufacturing, intelligent servitors, I think the only way we can hope to sustain an economy of any kind would be to redistribute the output of those machines significantly, or we would end up with a concentration of power around capital that utterly eliminates the economy as we presently understand it.
That would definitely be true in the case of self maintaining, self manufacturing, intelligent servitors who can do everything. I'm not sure we would get that, even as the end state. (We're probably at about the end state of "electricity can do everything", and yet there's still large amounts of manual labor. It can't do everything.)
> Ultimately, the production of pure automation springs directly from the resources of the earth, and the earth is everyones, so it makes sense to redistribute the production to a significant extent.
Unfortunately, under current law, the earth is not everyones. Real estate and mineral rights are pretty well entrenched. (For that matter, so are national governments. Given that resources are not evenly distributed, that matters.)
So getting to your philosophical starting point would require a massive transformation of existing human society. (Of course, robots doing everything might have that effect...)
> Ubi causing runaway inflation is another example of a thoroughly refuted claim that is easily given the lie by looking at extant UBI systems.
What are your examples of "extant UBI systems"? And to what degree are they true UBI?
> That would definitely be true in the case of self maintaining, self manufacturing, intelligent servitors who can do everything.
I think we will see enough of “everything” to invalidate the current paradigm of economic function. Something like 80 percent of all activity that does not involve high-touch customer interaction where people will prefer the more personal feeling of another human over the superior technical performance of a robot.
>Unfortunately, under current law, the earth is not everyone’s. Real estate and mineral rights are pretty well entrenched.
This is true, but we also tax those holdings in acknowledgement of their communal nature. Ideally, ownership would revert to the best steward, but we all know how that works out in practice lol. The status quo is only possible because the state retains a monopoly of coercive force. If the economic model is undermined, this monopoly is among the first of casualties. So while I don’t disagree completely, I’d have to say that we end up in a sort of race condition problem if the state hesitates too long to assert communal right to autonomous production.
> What are your examples of "extant UBI systems"? And to what degree are they true UBI?
Unfortunately all “UBI” experiments or implementations are limited in scope, because there is always a time or geographical limit to their application. So “true” UBI, which would largely eliminate the leverage that increased local wealth would have against market forces, has never been tried. As a result, existing trials have possibly been contaminated because the gradient at the edge is a problem.
Additionally, UBI in our context would be applied in an economic desert scenario, which means that wealth would not be increasing in the mean.
As for examples, there have been a few, but I’m on the can staring at my phone and I’m way too lazy to go back and try to search up the several examples of UBI experiments, so I’ll just leave you with the one I’m most familiar with because I lived there; Alaska’s permanent fund dividend. At 1-2k per year per human, it is a pretty big infusion for many families. Once again I’m too lazy to dig right now, but there has been many investigations of it’s effects on the economy, and the net result was that prices went -down- dramatically in response to dividends in an effort to capture market share. I would not expect to see this effect though, in the scenario we are exploring - once again it’s a gradient effect, I believe.
> massive transformation of existing human society.
something like a 3rd industrial revolution? As someone who is innovating in the developing technology in the sector, I believe that we are in for a much more sudden and extreme shift than both of the previous Industrial revolutions combined, compressed into a couple of decades. The first two spanned about 120 years of disruptive change between the two of them.
Yes, 1x Vexriscv RV32-IMAC + MMU, and 4x PicoRV32's as RV32E-MC for I/O processing, configured with extensions to enable deterministic, real-time bit-banging without having to count clocks.
It's a good model for MCU stuff. There were people pushing Chip Gracey (Parallax) to use RISC-V instead of his custom ISA when he designed the P2 a few years ago, but he chose to do his own thing. Which has made compiler development difficult.
This seems more on the RPI side rather than propeller, propeller was never a really good choice for production integration. This looks like it could hold its own in many contexts.
If I understand the architecture it's both -- a main MPU style core and then a bunch of PicoRiscV cores doing MCU tasks. The smart thing about using RISC-V here being having a unified ISA so you can compile programs that run on both or move between both, etc.
I'm assuming he probably has some sort of roundrobin shared memory access similar to what Chip did with "HUB Ram" on the P2.
The community is not the license. The “open source” development community is a user of that kind of licensing.
You might better describe them as the open source maintainer community. I do see how ai impacts maintainers. But I’ve dumped hundreds of thousands of loc into the bucket with no hope that anyone would really maintain it. With AI it might become part of something useful. The license has many uses.
I live in a developing country. What I find is that the corruption is generally easier to navigate here that it was in the USA. The corruption in the USA is much more entrenched, in the form of regulatory capture. At the local level this can look like a local ordinance where “only a contractor with xy and z (only one of which is needed for the job) can bid, favoring a specific contractor. Here you just figure out compliance with the person in charge.
> much more entrenched, in the form of regulatory capture
I am unsure how to interpret this comment. It is so broad that it dilutes the effect. Are there any wealthy countries that you feel do not suffer from the same issue?
App development: 4x. Lots of interface.
Firmware development c++: 10x+, with a huge increase in code quality. This requires extremely methodical planning, where planning and specification is 80 percent or more of the workflow…. But the results are outstanding. The “trick” is that planning and specification is basically code generation. By the time something is fully specified it’s basically done. The LLM just allows the humans to work in a more productive layer of abstraction and also offloads a whole lot of burdensome calculation and research.
Backend development: 10x at the prototype level it’s basically 100x magic. Everything one-shots if you go with simple python based cli tools with tests. Then you string them together to make the workflow. Once it works, you go over each tool with more testing and optimisation, library deletions where practical, etc. By building each significant function as a separate cli tool it makes testing/fuzzing automatable, simple and reliable.
reply