jj4211

Just two months ago, a security team member dinged one of our services for using Lets Encrypt, as "it's not as secure as a traditional CA".

I've had limited experience with slack, but the whole way conversations map to workspaces at least got to be confusing to me, and I would have liked an experience based on me as a user, rather than having my user span workspaces and have to juggle them to figure out how to talk to whoever I'm supposed to talk to at the time.

For me "it just works" doesn't ring true. Generally at least once a day, I join a call and it won't let me unmute, and I have to restart Teams.

Scrolling through history is obnoxiously slow.

The activity feed is mostly useless, spammed with stuff that isn't important and it's the only place that vaguely tries to keep track of 'Teams' conversations.

In my company, I've been added to about 70 Teams and it's pretty much impossible to interact with them, so as a result no one does, they all just start ad-hoc chats, since that's the only thing that vaguely gets managed in a way people can follow.

When going cross-organization, it's a crap shoot whether or not we can use text, voice, and screen share/remote control. I know this is generally due to obnoxious company 'security' policies and other solutions have it, but it is a frustration. One recent call with a particularly screwed up company had us on two different meeting platforms at once as well as on an old fashioned conference call, because text was only allowed on one platform, screen share on another, and no audio was allowed on either (despite both supporting all three).

Sure, Teams suffers, in part, because like all corporate tools it connects you to generally dysfunctional work communities. However it broadly does have it's own annoyances.

I would say LLMs specifically are in that ball park. Things like machine vision have been boringly productive and relatively un hyped.

There's certainly some utility to LLMs, but it's hard to see through all the crazy over estimations and being shoved everywhere by grifters.

Well, Trump specifically may not try because the risk/reward isn't really good for him.

As it stands, he gets to declare an unambiguous "victory" where he won at life. He got to be president with ultimately a clean sweep of the swing states and the popular vote and served as many terms as he is allowed to serve. Thanks to the rules, he doesn't need to compete again, and he can stop even pretending to work after 4 years.

Meanwhile, a push to establish him as "dictator for life" might at best buy him another few years in office before his health will fail. Such an effort comes with high risk, of him going down in history as more of a "bad man", of personal risk for being targeted by violence.

Now JD Vance might be game to make a go of it, he's got decades left in the tank. Of course broadly speaking there's a balance of power, with those currently in power relatively comfortable knowing that the vote serves as a nice way to get pushed out of office before people get pissed enough to put you in real physical danger. Plenty of opportunities to be self-serving with a pretty safe retirement should things start going awry. Fanaticism can drive people to go further, but I would like to think a pragmatic person with a sense of self-interest can see the value in a peaceful voting out versus having those same millions of people losing their political voice.

I find it insane that with modern computing and displays, they still just render a vague check engine light despite being able to easily display the specifics.

They mostly didn't have OSDs, they instead had indecipherable 7-segment and some fixed elements like 'Su Mo Tu We Th Fr Sa', with 2 or 3 buttons. The younger Gen-X/older Millennials got their reputation as 'whiz kids' in part by handling those interfaces on behalf of their mystified parents.

One, the volume knob is far quicker to respond than the usual 'up/down' slow volume adjustment on the wheel. The turn down the overly loud sound from the last driver immediately is nicer with a volume knob.

But with my car with hard A/C controls, I just reach down to the little 'up/down' toggle and tug it down a bit if I feel a little warm or bump it up a little if I feel too cold, or hit the big old button if I need to toggle it off to talk on speaker.

There are a fairly well known set of very common controls that will never be better and need an update. Coarse A/C adjustments, vent direction volume, and next-track are all no-brainers (unless you are Tesla...)

For example, here's a layout that obviously has room and depends on touch for a lot of features, but preserves a reasonably sane set of audio and climate controls (and four miscellaneous functions)

With that you don't look, you know pretty much immediately for the functions you would use.

There's still plenty of room for touch/voice controls for those more nuanced/complicated things that don't fit into button land well. Entering a navigation destination, managing any software updates, setting parameters like "should the car adjust cruise control based on speed limit signs, and if so, what adjustment to the limit should be applied?'

Particularly given the trend of 'glue a tablet to the middle of the dashboard'. If you are going to do that anyway, bring up a modern successor to the DIN/Double DIN standard, where the mounting is standard and update to also include USB-C for standard power, audio, and data. Add some network profiles for standardized exchange of useful information (Car speedometer, car model, fuel/battery amount and efficiency profile, navigation information to drive dash/HUD, etc).

To the extent you are able to (particularly if trying to stay legal).

So for streaming content, much of that isn't available to 'buy' at all. Even for the stuff you can "buy", technically speaking in many jurisdictions it's not legal to be able to rip your DVD or Blu Rays or remove DRM from a digital download.

For certain software, on-premise editions have been abolished or priced into the stratosphere because they don't want that market to exist anymore. Some of that software has competent alternatives, but sometimes your choice is dictated by your clients and partners, and opting for a less compatible or merely perceived as less compatible option is a non starter. Even among on-premise editions, a lot of software vendors have switched to still having it by subscription as the only legal way to keep using it. Again, maybe for those software you can get away by breaking the law as a workaround, but legally...

This is of course assuming the conversation narrowly applies to software type things. Everyone is also rebranding 'leasing' as 'as a service' and are copying much of the software playbook, for the same reasons, including making purchase of equipment more expensive to steer people toward the 'as a service' revenue strategy.

Then going beyond the 'tech' industry, it's getting really hard to buy a house rather than rent it from some company that has been pouring money into acquiring all the available real estate.

Problem is the data is rigged. It's road miles driven that autopilot deigned to activate for with cars that rarely need their friction brakes that are less than 10 years old versus total population of cars with more age and more brake wear and when autopilot says 'nope, too dangerous for me', the human still drives.

The other problem is people are thinking they can ignore their cars operation, because of all the rhetoric. A human might have still hit the deer, but he would have at least applied brakes.

Finally, we shouldn't settle for 'no worse than human' when we have more advanced sensors available, and we should call out Tesla for explicitly declaring 'vision only' when we already know other sensors can see things cameras cannot.

People drive drunk, people drive while checking their phone,

And those people are breaking the law.

people panic and freeze

I don't think I've ever seen someone panic so much they just act as if they didn't even hit a deer.

deers often just jump in front of you from out of nowhere.

In this case, the deer was just sitting there, so not applicable.

People hit fucking humans without braking because they’re not paying attention to what the fuck they’re doing!

If it was this much negligence, they'd be facing vehicular manslaughter charges.

But for some reason if it’s a car with assistance well now that’s scandalous!

It's scandalous when a human does it too. We should do better than human anyway, and we can identify a number of deliberate decisions that exacerbate this problem that could be addressed, e.g. mitigation through LIDAR, which Tesla has famously rejected.