We're building something along these lines, but since our roots are a consulting business, we're still building around the idea that there needs to be an expert integrator doing the front-loading work of discovery/decomposition/scoring of tasks/implementing them as those agents. These tools are terrifying to anyone not quite technical, and it turns out, people are bad at decomposing their own work, let alone describing it in a box with a blinking cursor.
We're obviously going to be holding ourselves back in terms of scale and in terms of not being a "true" SaaS with this approach, but my thesis is that we get much higher quality results and higher compliance/activation and can charge more for the bespoke model backed by our own platform.
This is the craziest part. We've long had publicly subsidized private projects, or corporate tax breaks given to entities with the fig leaf explanation that the projects will bring economic activity, or jobs, or some sort of long lasting, durable benefit, or even at its most craven, keep a stupid populace happy that their favorite sportsball team hasn't left their town.
Datacenters do none of these things. They don't bring any true employment numbers to make a difference. They don't materially improve their surroundings. They don't increase land value. They aren't an attractive neighbor. They don't benefit the tax base, not least of all when you're doling out massive tax breaks and deferments... no one _wants_ them, so it's not even like this is a "well, you're just not a Buffalo Bills fan, but the rest of us are, get on board!" situation.
There is zero net benefit to the citizens in these places to welcome these facilities to their town. Shorter latency to their chatbot of choice, maybe? But in any practical, moral hazard sense, these are all pure net negatives for the communities, and it's wild that these leaders think they're some sort of marquee, glamorous, prestigious win of a project.
Compare this with other, things-that-rhyme-but-aren't-the-same projects, like the TSMC fabs in Phoenix: these projects are bringing a ton of high-paying, new jobs (and, somewhat controversially, an expat community from Taiwan to help onboard them in the meantime), but they're also delivering in other economic terms because of the supply chain's knock-on effects: the TSMC fabs further the reputation of Phoenix as the Silicon Desert that Intel, OnSemi, Microchip, and Motorola had long been working towards, but at a much more amplified scale, and in a truly meaningful capacity. The money being spent here is staying here, and driving some real practical benefits. But even still, it's an open conversation around how careful we need to be with the water usage of these fabs (though TSMC is aiming towards 90%+ recapture in the next few years, I think it's ~60% right now), and other considerations... they are still, on balance, bringing 6,000-12,000 direct jobs, and even more indirect jobs as they continue to expand.
These datacenter projects don't even do _that_ well. They're just upsetting the power grid and creating unfortunate microclimates for the immediate vicinity for a handful of NOC jobs. (And some itinerant construction and engineering jobs.)
> it's an open conversation around how careful we need to be with the water usage
It's only an "open conversation" because of the complete lack of transparency electricity companies and data center companies are required to give us. Communities are expected to make these decisions without any of the data to know what they're signing up for.
What irks me the most is how prevalent the industry line about water usage "not being a big deal" is on HN. Closed loop systems are probably less than 10% of data centers! Not to mention the fact that even in closed loop systems, you still have to regularly "bleed the lines". That water builds up with the anti-freeze, anti-corrosives, and anti-fungals additives as well as other stuff they pick up along the way including PFAS. Closed loop systems are far from a perfect alternative to the open loop systems that are prevalent. They can use as much as 40% more electricity, have much higher upfront costs, and then you have to have a solution for the toxic sludge problem.
This article in the WSJ was by an author who decided to induce immunity to poison oak by exposure: eating it. It's... quite the journey, evidently, but possible:
Right, but there's a core conceit we use in the US (mostly) that you are innocent until you are proven guilty, and if you are wrongfully accused (as was evidently the case from the author), you should perhaps NOT be put into such a grim set of living conditions with essentially no rights.
In this case, the author evidently _was_ a law abiding person, so the optimization failed, senselessly, likely out of a systemic effort to strike enough fear in the populace to over-index towards avoiding the possibility of this sort of situation. (Much like Singapore caning people for minor offenses.)
Whether or not you agree that such draconian punishments or processes are effective or fair is a different discussion, but this person was LITERALLY not supposed to be in jail, so how fair is it that they were removed from polite society for over a month in such poor conditions and at considerable expense?
They _do_ specifically protest, and it's crazy that they're able to detain you like this from an accusation while they build a case, even if you're innocent. In the US, barring flight risks and past history or cases of real malice or violence or an ongoing threat, you can at least typically make bail, AND the conditions in a jail are generally far better and less strict than this:
>Both cases were ultimately dropped and the second arrest was essentially tied to the first and shouldn’t have even been possible. But because of how the system works weather it’s a viable reason or not, they can still trap you in there for a time while the case is being reviewed. I met others who where there for shorter and much longer periods of time. The worst part was knowing i was innocent. After it’s all said and done you walk out and they act as if nothing happened. Not only was this was all extremely traumatizing but it cost me a HUGE of money that I really did not have and caused irreversible damage to my life.
> In the US, barring flight risks and past history or cases of real malice or violence or an ongoing threat, you can at least typically make bail
The literal majority of people in US jails are there not because they have been convicted of anything but because they were given a bail amount they couldn’t afford to pay, which is a deliberate strategy by the courts when there is no justification to refuse bail. This can look like a $500 cash bail set on a homeless guy charged with resisting arrest (aka being arrested). Many of them are innocent and are trapped and have their lives ruined in exactly the way this guy describes. (We assume that many of them are innocent because when someone pays their bail, more than 50% of cases are simply dismissed as soon as they leave jail. The expectation is that they will just plead guilty because otherwise they are stuck in jail for months waiting for a trial).
> We assume that many of them are innocent because when someone pays their bail, more than 50% of cases are simply dismissed as soon as they leave jail.
The legal system doesn't have the resources to move forward with the case and decides it isn't a priority. I've seen this happen many times with people I know committing violent felonies.
Even for smaller examples it happens all the time. Half the time you can completely get out of traffic tickets by showing up to court to plead not guilty. They dismiss the case because it's not worth the time.
> They dismiss the case because it's not worth the time.
I don't know what this means in the context of the US justice system. They're not paid on commission. They're being paid to be there no matter what happens.
They dismiss the case because the cop didn't bother to show up, or they didn't have any evidence against your defense. The reason you (as the person who got ticketed) don't show up to court is because you know you have nothing to say, or because it's not worth it to you when getting out of the ticket isn't enough pay for 3-4 hours of your time. The only reason you do show up is because you think you have a defense.
If you can't make bail, you're showing up no matter how stupid the charge is.
edit: I have personal experience (from a few decades ago) of being forced to face stupid charges. It was a game. They inflated the potential sentence to 3-5 years through silly charges designed for just that, and offered me a plea bargain of no time, no fine, and expungement from my record in 6 months. I pled guilty. If I hadn't been bailed out, I would have had to wait two weeks in jail for that moronic, depressing event. I pled guilty because it was easy to do, even if I hadn't done anything. If I had sat in jail for two weeks, I might have pled guilty even if it involved a week of jail time and a fine, just to get out.
Kalief Browder spent almost 3 years in Riker's Island awaiting trial just to have the charges dropped. People on here told me that showed that the justice system worked. I said that his life was destroyed by this, and he would probably end up dead soon. I got downvoted furiously. He'd killed himself 2 years later.
Large Criminal Justice systems like NYC have a large population and it's easy to end up being thrown in a cell and forgotten. Having a lawyer or not is the biggest difference in outcomes. I sat in on a lot of court cases in Philly when dealing with a case. Saw 17 year olds locked up with no lawyer over a simple drug case while a guy caught dealing pounds hired a lawyer and got off with nothing after completing a "rehab" program. Guy didn't even use.
Dropping legitimate cases due to priorities and resources doesn't mean that they don't also still often pursue illegitimate cases beyond the point of reason.
I've actually thought of a much more dystopian idea: that Waymos could be technically used as roving traffic cameras, and report on the human drivers around them. They absolutely have strong enough telemetry systems to be able to determine things like excessive speeding, dangerous lane changes, red light running, etc., and their imaging systems could probably pick up a license plate with little additional modification... it's obviously not great from the perspective of general optics and morale, but it would surprise me if no one had floated basically WayNarc as a business model...
Any time I've loaded something into the trunk of a Waymo, it pre-emptively pops the trunk when I'm getting out _for_ me and reminds me to get my things from the trunk, so this is... surprising as a failure mode. Wondering if there was some issue with the latch/opening system, because it's definitely programmed to work the right way. (Or he tossed his stuff into the trunk from the main cabin, but... it's a pretty low hatch ceiling there.)
None of that is the problem, shit happen, the problem is them asking them to come get their stuff instead of apologizing and sending the luggage at their own expense with a free ride at a future time.
> My guess would be the Jaguar's CAN bus being the weak link
Puzzling because trunk open sensor is already a thing and making sure it’s triggered after a ‘open trunk’ command is issued is not exactly rocket science:)
Besides the CAN bus there's also a lot of other steps along the chain, their in-vehicle comms network, the uplink they're working with
Waymo's own systems are sending the current volume and HVAC settings to the OEM CAN bus every tick (presumably because the HMI's traffic is very low priority and may not be delivered on their network), maybe the trunk release doesn't tolerate repeat activations because of some quirk?
At that point obviously I'm guessing a bit, but the fact this issue has been prominent for so long makes me think there's some platform-specific quirk that's making it more complicated than it looks from the outside
Not that surprising if the thing that failed was the thing that notices whether or not you put something in the trunk in the first place. Unless it does that routine at the end of every ride, regardless of whether it thinks something is in the trunk or not, then it's not a fail safe system and occasional mishaps like this should be expected at scale.
We're obviously going to be holding ourselves back in terms of scale and in terms of not being a "true" SaaS with this approach, but my thesis is that we get much higher quality results and higher compliance/activation and can charge more for the bespoke model backed by our own platform.