I would consider a lot of mobile apps to also be a 'micro-payment' type model. Clearly there's no issue with people paying for content, I think the real gap here is in the ability for the consumer to pay for the content. If I go to some random news site and it hits me with a paywall for a micro-payment there isn't a simple system by which I can actually give them money without directly signing up for a subscription to that specific site or some other service. If there was a type of wallet for this that I could just put money into and sites asked "would you like to pay X amount from your wallet to read this content?" I would be more amenable to it. It's the same idea with streaming sites and piracy. Companies have made content more expensive and more exclusive so why would I want to jump through the extra hurdles which was supposed to make consuming your content EASIER. It's always about ease of access to the consumer.
Based upon everything we know from Valve's corporate structure, they're basically their own self-contained YCombinator. They have a ton of internal 'startup' groups that are constantly trying new and boundary pushing ideas. Looking at it through this lens, it's not really a good company fit for any junior. Having juniors detracts from your seniors work, but the point is that you're supposed to get value from that when they eventually become mid-level or senior engineers themselves. But if you're constantly working in new complex environments it's hard to bring a junior up to speed and teach them the requisite skills to thrive, especially if that project they just spent 3 months getting up to speed on gets canned because the idea didn't actually pan out.
I was worried about this when I turned it on myself, but under the usage panel it shows that it limited my spending to just the $50 and that auto-reload is off, so it doesn't seem this would be the case.
I turned on this overspend and limited the spending to $20. A day later I checked my spending, I had used "295%" of my limit. Almost $60. No idea why it didn't respect my setting.
I've been using Gas Town a decent bit since it was released. I'd agree with you that it's design is sub-optimal, but I believe that's more due to the way the actual agents/harnesses have been designed as opposed to optimal software design. The problem you often run into is that agents will sometimes hang thinking they need human input for a problem they are on, or they think they're at a natural stopping point. If you're trying to do fully orchestrated agentic coding where you don't look at the code at all (putting aside whether that's good or not for a second) then this is sub-optimal behavior, and so these extra roles have been designed to 'keep the machine going' as it were.
Often times if I'm only working on a single project or focus, then I'm not using most of those roles at all and it's as you describe, one agent divvying out tasks to other agents and compiling reports about them. But due to the fact that my velocity with this type of coding is now based on how fast I can tell that agent what I want, I'm often working on 3 or 4 projects simultaneously, and Gas Town provides the perfect orchestration framework for doing this.
the problem with gastown is it tries to use agents for supervision when it should be possible to use much simpler and deterministic approaches to supervision, and also being a lot more token efficient
I strongly believe we will need both agentic and deterministic approaches. Agentic to catch edge cases & the like, deterministic as those problems (along with the simpler ones early on) are continually turned into hard coded solutions to the maximum extent possible.
Ideally you could eventually remove the agentic supervisor. But for some cases you would want to keep it around, or at least a smaller model which suffices.
This press piece is specifically about their client software for streaming. The actual host PC is likely a Windows PC specifically for these kinds of compatibility purposes I'd imagine.
Some time ago, The server side of GeForce Now utilizes VMs running on servers with partitioned Quadro cards, with each partition assigned to one client VM running prepared Windows machine
It's likely then that you are thinking too small. Sure for one off tasks and small implementations, a single prompt might save you 20-30 mins. But when you're building an entire library/service/software in 3 days that normally would have taken you by hand 30 days. Then the real limitation comes down to how fast you can get your design into a structured format. As this article describes.
still seems slow! I’m saying what happens in 2028 when your entire project is 5-10 minutes of total agent runtime - time actually spent writing code and implementing your plan? Trying to parallelize 10m of work with a “town” of agents seems like unnecessary complexity.
I think that most of the anecdotal and research experiences I've seen for AI agent use so far tell us that you need at least a couple pass-throughs to converge upon a good solution, so even in your future vision where we have models 5x as good as now, I'll still need at least a few agents to ensure I arrive at a good solution. By this I specifically mean a working implementation of the design, not an incorrect assumption of the design which leads the AI off on the wrong path which I feel like is the main issue I keep hearing over and over. So coming back to your point, assuming we can have the 'perfect' design document which lays out everything, yeah we'll probably only need like 5 agents total to actually build it in a few years.
Isn't this the case for almost every product ever? Company makes product -> markets as widely as possible -> only niche group become power users/find market fit. I don't see a problem with this. Marketing doesn't always have to tell the full story, sometimes the reality of your products capabilities and what the people giving you money want aren't always aligned.
This article is such a weird take in my opinion and I think it actively proves why our current media environment is harmful. It seems to be making the assertion that people are over reacting about microplastics due to the challenge made in the paper about the way these things were being measured. Note: no where does this paper challenge that microplastics are 'actually' harmful, only that we aren't measuring that properly. To then extrapolate that to "Our stance towards modern technology is toxic" is kind of wild considering we just came out of the lead gas era...
That site is a political advocacy org for a certain brand of economic liberalism. At least they're pretty open about it:
> The Argument is a mission‑driven media company based in Washington, D.C. We make a positive, combative case for liberalism through sharp, well-argued opinion pieces, original reporting, and multimedia content that confronts the illiberal drift in our politics.
We aim to persuade, not preach; argue, not just diagnose. Our coverage will focus on the politics and economics of growth, technology and society, gender and family.
The biggest thing for me is I don't understand how people can sleep with these watches on, it's so uncomfortable to me personally which is why the different ring technologies appeal to me more. I just wish either Garmin made one or that there was one I didn't have to buy a subscription to use.
When I started wearing a watch, it constantly distracted me and cause minor annoyance all day while I got used to it. At some point it just started blending into the background. Now, when I wear my watch on my other hand, I still have the same problem, but I don't notice my watch normally anymore.
I've slept with my watch for a while (stopped because the battery is crap and I need to charge it every day or it won't last past noon the next day) and I've had the same adjustment period.
I also think some people are just more sensitive to these things. Some sleep with full-on CPAP machines hooked up to their faces like it's nothing, but others can barely stand wearing clothes when they're sleeping. Plus, some watches are more comfortable than others.
> I've slept with my watch for a while (stopped because the battery is crap and I need to charge it every day or it won't last past noon the next day) and I've had the same adjustment period.
I'm guessing this is an Apple Watch? Garmins typically last a week or more.
Samsung Galaxy Watch from years ago, still running Tizen. Got it as a company Christmas present, so I'm not too fussed about the mediocre battery life. From what I can tell the Garmin models in the same price range don't last much longer either.
If it makes you feel any better, I've got a perhaps Galaxy Watch 7 from almost a year ago or so running wearOS, and it's battery life is at best at the 36 hour mark regardless. Came from an older tizen-based watch with similar battery life, and at least wearOS is far more funtional.
Not sure why Garmin or any of the exercise tracking watches are being used for sleep tracking. They're infamously bad at it from my experience.
The rings (notably Oura) are much better. I used to wear both and they gave completely different results, with the Oura being far more accurate to how I feel and the timings of going to sleep and waking up. Garmin almost always reckoned I woke up an hour earlier than I did and ended the tracking there.
It's honestly best not to get too involved in tracking sleep. The analysis does more to ruin your mood and give you nocebo effect than it really gives useful information.
I will confess, I do still wear my Garmin to bed because I like the vibration alarm over anything audible.
Surprisingly, I find the Apple Watch significantly more comfortable and less perceptible than any other watch I've worn in the past. My theory is it's the contoured back for the heart rate monitor, etc. But maybe I'm just weird.
I think it depends what kind of smartwatch you choose. It should fit your wrist. The amount of people I've seen with e.g. a Galaxy Ultra watch that stands over their wrist on both sides is staggering to me.
Nylon straps are pretty comfortable. The default strap with the metal buckle would dig in to my wrist, and also it was difficult to have a perfect fit since you had to pick a hole. The nylon strap allows for analogue refinement.
I have a light Huawei ("Honor") Magicwatch 2 for €45 with an old apk, and it's much better than my son's expensive Garmin. I got a good leather bracelet from a €15 Chinese smartwatch. Very comfortable. Battery lasts a week.
I don't wear one anymore, but I used to put in on an ankle overnight. I would wear a sock on that foot so that I didn't tear it off with my other foot during sleep.
>does feel its back to might is right, and the last 80 years of relative peaceful times is sunsetting.
Depending on your perspective, 'might is right' never changed. The US has forced its policies on other nations through quiet force for a long time. I think the only thing that's changed is that Trump wants to say the quiet part out loud now which makes it way easier to push back on. Combine that with the fact that Trump has 0 political ambitions outside of just being in power and it becomes very easy to just ignore what you hear coming from the top. Often it clearly has no thought put behind it, seems vindictive in nature, and is forgotten the next day, like a child's tantrum. To circle back a little, now that the US in such a passive state due to this, a lot of other countries feel safer to push their influence on the world because they see no repercussions for what others are doing.
reply