Perhaps we need an AI based company that scans all code for security issues, and fixes it on the fly?

(/s... of course.)

This is what the US did in ages past. Invested in state of the art stuff, like electricity, plumbing, water/wastewater, etc. Texas had much of the state redone around lakes to deal with flooding, as well as provide a stable supply of fresh water. They invested... and the ROI was awesome, making towns like Austin actually doable.

You would never see that now. At best, money goes to private companies for stock buybacks.

If one compares the US to a farmer, the US isn't doing any tilling or dropping seed, and then wonders why there is no harvest to mention, all the while, China and other nations are actively seeding, fertilizing, etc.

Seems the UAE is starting to do do long term investments to ensure they have a future, post-oil.

Refactoring? Now that is a word that in many "Agile" environments, I've never heard mentioned, with management saying, "we will cross that bridge when we come to it", not caring about the mountains of tech debt, but getting mad and firing people when they slam into some tech debt which slows them down that the PMs and marketing say that they already sold those features to customers, why are they not in yet.

What really needs done is to have a "thread" of refactoring going on in the code base. Sprints need to actually end and have time in between them. If a Scrum master says that everyone should be in a sprint 24/7, sack them, as they are just a mouthpiece for marketing. Additionally, every 12-18 months, have a feature freeze of 3-6 months to just do a complete refactor, maybe even rewriting into a better language, like moving from Python to Rust for server side stuff, which generally will help performance. Having that bug hunt will help things overall, and give a consistent base to go from there for the next feature additions.

Ultimately, it is about how teams work. Methodologies are important as well. Waterfall has gotten us to the moon. Agile seems to have gotten us the next rev of Windows.

Like Scrum is bolted onto Agile, we are likely to see vibe coding bolted onto Agile as well, with workflows of AI creating stuff, it going through debug cycles, tested, staged, and then thrown into production... hopefully all in different environments.

I've wound up moving to OPNSense for my main firewall, from PFSense, although these are on dedicated devices. Overall, in theory, it might be better to use a BSD because it has a different set of bugs than Linux, so something that is a kernel level issue would be stopped by that.

I find it a hard choice, if I were choosing a NAS OS. The main reason why I have wound up going for Linux in general, is the support and tools available. Technically for a NAS OS, maybe FreeBSD is best with their ZFS, perhaps even Solaris, but I've been running ZFS on Ubuntu for years without issue, using snapshots for backups [1].

[1]: I pop a snapshot, then have Borg or Restic back up that mounted snapshot directory, as opposed to using binary images, just to make restores faster.

I'm going to go about splitting hairs here.

UNIX, as in official, trademarked UNIX flavors are macOS, AIX, Solaris, and HP-UX. Technically, others are unix-like operating systems. However, these days, this is just historical trivia. The trademarks and certifications are expensive, and tend to matter for bean counters.

FreeBSD, NetBSD, and OpenBSD are splits from Jolitz's 386BSD... not to be confused with BSDI's BSD/386. These are "true" BSDs, from the 4.4 BSD Lite code base, after the lawsuit that separated SVR4 from BSD.

Linux, on the other hand is mainly GNU. Originally, a lot of Minix stuff was ported to Linux to get it going, including its filesystem (which was replaced by xiafs, then extfs, then finally ext2 [1].) It is a little bit more AT&T like than BSD... but these days, those differences don't matter, as both BSD and Linux use ps commands that support both SVR4 options (ps -ef) and BSD options (ps aux). Linux is a completely different development effort than BSD.

There are different philosophies at work as well. BSD tends to use chrooted jails, which are effective. Linux has that option, but has gone with containers and virtualization as other options.

As for what is better, that debate has raged on for decades... Just use what is right for the application at hand.

[1]: You might even be able to use umsdos... which gave you UNIX permissions, long file names, and case sensitivity, if you felt insane.

This is what I did. What people need to consider is writing a functional replacement to systemd, perhaps split its functionality up into smaller modules with very detailed security in AppArmor and SELinux, so if one module gets compromised, the damage is tough to spread.

However, for now, it is there, and has a few nice features, so might as well move on.

I still wonder if systemd's code has ever been audited though, especially recently. Just seems ripe for a remote root hole somewhere in that huge mass of code.

There is a balance somewhere.

On one end is that someone, somewhere, will take grave offense to a comment, no matter what. On the other end is exactly as the parent mentions -- we are back to trolls and people cursing like the old school XBox days.

What might be useful is limiting the size of an audience, and having posts disappear (archived, removed) after a while. That way, there is reputation to be dealt with, but it isn't hoping that one's post passes every single psychopath in the world.

If I had a choice between government and business, it would be a tough choice. At least government tends to not sell the info it slurps up on its citizens, and because business sells the info to government for law enforcement reasons anyway, might as well go with only one layer of surveillance. Government also generally doesn't have a profit motive, which ensures some level.

Plus, governments can be shamed if they have a leak. Businesses can lose all their data to hackers, and if they do some PR tap-dancing, their stock will be back and climbing in at most 1-2 quarters. Governments will lose face and actually try to prevent stuff like this from happening.

Then you will get people not bothering using it, for fear that someone will set their car or house on fire if they said the wrong thing politically.

What we might need is something that is both. A pseudo-anonymous network. Real info, but allows for multiple identities. If reported or something like that, then action can be taken against all accounts linked to it. This way, someone who gets someone after them can nuke that identity and use another... but ultimately, they are responsible for their stuff, regardless... but it keeps the stalkers and trolls at bay. Of course, ID theft can be an issue here, but nothing is perfect.

There is one nice thing about EVs. They are damn quick. They can be fast, but the quickness is nice, and in urban areas, you don't need top end. You need quick, to get on the highway, to get around obstacles, to go from 60-0 when someone tries a "crash for cash" on you, or gets in a lane ahead, panics because cars ahead are stopping, etc. They drive very smoothly, and the fact they are so quiet is a blessing. They are not for everyone, but they are "good enough" for many people.

Another thing is that some EVs can provider inverter power. This can be useful on camping trips, although not as useful as if it were on a PHEV or a serial hybrid that can fire up the range extender.

Right now, the bottleneck are chargers. I'm lucky to have a 50 amp, 240 volt outlet outside, so throwing a level 2 charger outside is a matter of plug and play. However, very few people in urban areas have access to parking, much less a dedicated charger, and racks of dedicated chargers require a lot of juice from the power company... juice which the Bitcoin mining companies have first dibs on (to the point where mining companies get paid by power companies not to mine.) Because of this, I'm looking at PHEVs and serial hybrids for my next vehicle, especially with the local grid not being that reliable, especially in winter or if a hurricane comes by.

The Volt was something that addressed the range anxiety issue in a decent manner, with a usable range extender. Charge it, take it to a gas station and fill it up... who cares. One could just use it as a regular car if one didn't have a dedicated charger, or an EV with one. Best of all worlds.

GM killed it.

The only thing similar are bespoke Edison retrofits and the RAMCharger, which isn't even in production yet (and weighs ~7500 pounds.) It would be nice to see more serial hybrids.

Probably the best solution right now are PHEVs. Wish more models were available, like the Jeep Wrangler 4xe, Prius Prime, etc. They may not have much in the way of battery range, but if someone's commute is 10-20 miles, the smaller battery might just be good enough, especially if there is a home charger available.

We may be a bit closer... Microsoft has stuff going in this arena. Will it actually pan out? Who knows. All the while China is waiting in the wings with what they make.

Once someone is able to factor ECC algorithms and has the ability to put bogus transactions on blockchains, say buh-bye to cryptocurrencies as we know it.

It isn't "there" with desktops and servers yet, but it is coming along, and can help add features to subsystems.

For example, having a RAID card that has an onboard DRAM cache, but upgrade it to use ZFS on the backend, even though it will present only as zVols, one would get the benefits of ZFS, including better RAID management, resilvering, dRAID and other items, as well as have the ability to have FDE moved to the array controller, which means no worries about compromised data from a drive pulled from an array. Or for other SBCs, more features that can be added, or added security, such as containers.