I feel like the cloud hosting companies have lost the plot. "They can provide better uptime than us" is the entire rationale that a lot of small companies have when choosing to run everything in the cloud.
If they cost more AND they're less reliable, what exactly is the reason to not self host?
> If they cost more AND they're less reliable, what exactly is the reason to not self host?
Shifting liability. You're paying someone else for it to be their problem, and if everyone does it, no one will take flak for continuing to do so. What is the average tenure of a CIO or decision maker electing to move to or remain at a cloud provider? This is why you get picked to talk on stage at cloud provider conferences.
(have been in the meetings where these decisions are made)
Plus, when you self-host, you can likely fix the issue yourself in a couple of hours max, instead of waiting indefinitely for a fix or support that might never come.
I mean, I still prefer to have the ability to fix it myself, because I know I can probably do it in 1h max. I know this doesn't apply to most people, especially those outside of HN though.
Even if resolution times are equal, there is some comfort in being able to see the problem and make progress on it to feel like you're actively doing something. I work in a large enterprise and we have a team dedicated to managing critical incidents and getting everyone together for a resolution. When a 3rd party vendor is the reason for the outage, those calls are really awkward. It's a bunch of people sitting around pressing F5, all frantically trying to make it look like they are actively helping, when no one is actually doing anything, because they can't.
I equate it to driving. I'd rather be moving at a normal speed on side streets than sitting in traffic on the expressway, even if the expressway is technically faster.
Today a client is having some issue with Zoom because of some artificial rate limits they impose. Their support is not responding, the account can't be used, courses can not be held and there's not much we can do.
We already started looking into moving away from Zoom, I suggested self-hosting http://jitsi.org
Based on their docs, self-hosting is well supported, and probably a $50-$100 server is more than enough, so a lot cheaper than Zoom.
Artifical limits, because they have 40 paid licenses that they can not use, because of a non-disclosed assignment limit that is NOT mentioned in the pricing page nor in the ToS.
If a company doesn't respond to this it tells you they likely only respond to lawsuits. As a paying customer whose business operations are impacted, you should have standing to sue. Your company could potentially extract from Zoom the entirety of the money that their dumb decision made your company lose. Consult a lawyer for actual advice and next steps.
Of course, it's also possible you signed a contract that basically says "we can just decide not to work and you can't do anything about it" in which case, sucks, and fire whoever negotiates your B2B contracts. But also, those clauses can be void if the violation is serious enough.
Probably not worth the effort, for a couple days of downtime, we'll just move somewhere else.
But I agree, I recognize the silence in that forum thread that was locked without a resolution: some boss said "let they complain or pay, we don't care about them otherwise".
For a start-up it's much easier to just pay the Cloud tax than it is to hire people with the appropriate skill sets to manage hardware or to front the cost.
Larger companies on the other hand? Yeah, I don't see the reason to not self host.
TBF, it depends on the number of outages locally. In my area it is one outage every thunderstorm/snow storm, so unfortunately the up time of my laptop, even with the help of a large, portable battery charging station (which can charge multiple laptops at the same time), is not optimistic.
I sometimes fancy that I could just take cash, go into the wood, build a small solar array, collect & cleanse river water, and buy a starlink console.
Rust is one of the few languages where I found AI to be very well checked. The type system can enforce so many constraints that you do avoid lots of bugs, and the AI will get caught writing shit code.
Of course, vibe coding will always find a way to make something horribly broken but pretty.
I have noticed LLMs tend to generate very verbose code. What an average human might do in 10 LoC, LLMs will stretch that to 50-60 lines. Sometimes with comments on every line. That can make it hard to see those bugs.
Most people need a recommendation for something more current, from people who work on these modern cars daily. The reputation of 25+ year old models can be misleading.
Another source of good recommendations could be insurance companies. Cars with low reliability or very expensive fixes probably need more expensive insurance. But I don't know if this data is public or if you can tell apart the reliability from the repair cost.
If you're in Europe, you can consider Dacia. A lot of their stuff is old Renault parts that they've bought a license to use/manufacture. Get a pre-2023 model with the 1.6 non-turbo non-hybrid petrol engine - it's actually a Nissan HR16DE, which has been in use since 2004. Very reliable and low complexity.
Is it using that Nissan/Renault CVT? That transmission is notorious junk.
I must say that I've been impressed with Dacia. Even the build quality is excellent - on par or beating VW. I've driven on Romanian roads so I can see why they would prioritize such high build quality.
You can get them with a manual transmission, or a dual-clutch automatic, or CVT. AFAIK, the manuals are all decent, although the 6 speed manual on the 4WD models has quite low ratios (no transfer case) so it doesn't have great fuel economy at highway speeds.
Not true unless you equate a normal road (one or two lanes per direction, rarely separated from the other direction) with a highway, which is something else (comtrolled access, separated lanes, safety lane). Top highway speed is 130 kph, express road speed is 110 kph and normal roads it's 90 kph.
At least the last time I was there, 2010, all the intercity highways look like they hadn't been maintained since Ceausescu fell. No matter what the legal limit was, there were very few places where one could drive 90 kph safely. Maybe this has changed - I certainly hope so
Well, it's quite different now after being in the EU for nearly two decades. What you recall is regular roads, that cannot be called highways, but those got fixed as well. The highway infra is still not countinous, but it exists. And Romanian drivers do >90 kph on normal roads as well. Romania and Bulgaria have the highest road fatalities per capita in the EU.
It's not an infrastructure issue, but a cultural one that took off because of lacking infra. Those roads were designed for doing a maximum of 90 kph on them. Drivers were out of options, needlessly wasing time on thd road, so they started driving recklessly.
Pretty much every major safety feature is an order of magnitude less meaningful than the last.
If you wear a seatbelt and eschew the most risky driving behaviors your chances of getting in a crash where the difference between 2005 and 2025 matters are very, very, very, small.
At the very least, modern cars are much heavier and ultimately mass wins. For example, a 2005 Honda CRV weights 3400 lbs while a 2025 is 3900 lbs.
Plus they have tons more auxiliary safety features like lane departure warning, forward collision warning, blind spot detection, better visibility, etc. And they are roomier, have more power, get better gas mileage, and have backup cameras and Apple CarPlay!
Nothing I own has ESC. It's not even close to an issue. Maybe I'd feel differently if I had something with a ton of power but I don't.
None of these technologies prevent you from coming into a turn or stop or other situation too hot which is far and away the biggest problem in snow and is easily doable independent of vehicle equipment.
Sure if you got good tires, good weather, drive very cautiously, have long wheelbase vehicle - it will never be used.
Drive a bit faster on gravel road and FAFO. Puddle on a highway - hope you can brake quick enough to trigger ABS and save your ass. Taking a bend with a bit more speed than you should and there's some dirt - RIP. There's some snow and you drive Smart - you'll be spinning even with the best of tires.
Sure it can be avoided. It saved my ass about 5 times. 4 of which I was simply going too fast, but once it wasn't my fault.
I'd suggest getting car with one and try to spin it out of control either on gravel or snow. Pretty sobering.
Crash safety has become grossly exagerrated because the standards have been sharply rising last few years. Most 15yo cars will keep you safe just fine in a median crash.
A 15 year old car currently is going for 5 figures - not a shitbox. Not unless it’s a shelll of rust held together by bondo. Then your crash standards or whatever year are meaningless as the chassis may have 25% or less of its design strength.
At least over here where we have mandatory inspections you can find statistics on percentage of cars which fail the inspections, broken down by brand and model. Toyota seems to consistently place in the top.
Those sorts of comparisons are highly misleading because the overwhelming majority of failures for any inspection program are simple stuff that doesn't affect the operation of the vehicle in the base case. Light out, bald tires, brakes below replacement threshold, windshield crack, minor exhaust leak, etc. So what you wind up measuring by proxy is the owner behavior, since that's the dominant factor in how proactively those sorts of things get addressed.
And it ought to surprise nobody that trophy wives in 4runners show up with their vehicle in a statistically different state of repair than single moms in Altimas.
The big failures that you really want to avoid almost never show up on safety inspection data because they typically render the car much less drivable so they either get fixed promptly or the car stops coming around for it's inspection.
That's very interesting. I could argue that you are reading the signal wrong here. You want to go for the car that has the most failures in some cases, since it has survived long enough to fail in minor ways that leave it still able to drive.
If you have car brand A that has a reputation for having catastrophically expensive failures in major components, and car brand B which just keep chugging along for decades, you will probably see an elevated failure rate for brand B since it is still driving, while brand A will not be failing since it has already failed so badly it has been scrapped.
Sadly there's nothing in the license of seaweedFS that would stop the maintainer from pulling a minio -- and this time without breaking the (at least spirit of the) terms of the project's license.
reply