It feels like every few months there's a new tech "revolution" being hyped up as the future. Besides AI, what’s the most overhyped trend in tech right now? For me, it’s the constant buzz around the metaverse.
I hate this one. People who have been following carbon capture for years know who is doing it sincerely and who is using it for greenwashing. Of course oil companies who say they're doing carbon capture are doing greenwashing, no one should be surprised about that. Companies like Climeworks are doing real CCS.
Cloud. Businesses went all in on cloud under this illusion of stable costs, but costs go up and contol/support have gone down, and I'm seeing businesses spin on-prem back up.
1000% this. Without giving away too much information, I work(ed) for a cloud provider (not one of the big ones, there are a surprising number of smaller ones in the field you've probably never heard of before). I quit this week to take a position in local government with some quaint, on-prem setup.
We were always understaffed for what we promised. Two guys per shift and if one of us took vacation; oops, lol. No extra coverage, just deal.
Everyone was super smart but we didn't have time to work the tickets. Between crashes, outages, maintenance, and horrendous tickets that took way too much work to dig into, there was just never enough time. If you had a serious problem that took lengthy troubleshooting, good luck!
We over-promised on support we could provide, often taking tickets that were outside of infrastructure scope (guest OS shit, you broke your own server, what do you want me to do about it?) and working them anyway to please the customer or forwarding them directly to one of our vendors and chaining their support until they caught wise and often pushed back.
AI is going to ruin Support. To be clear, there will always be support and escalation engineers who have to work real problems outside the scope of AI. However without naming names, there's a big push (it'll be everyone before too long, mark it) for FREE tier support to only chat with AI bots. If you need to talk to a real human being, you gotta start dishing out that enterprise cash.
Mix all that together and then put the remaining pressure on the human aspect still holding things up and there's a collapse coming. Once businesses get so big they're no longer "obligated" to provide support, they'll start charging you for it. This has always been a thing of course, anyone who's worked enterprise agreements knows that. But in classic corpo values, they're closing the gap. Pay more for support, get less in return. They'll keep turning that dial until something breaks catastrophically, that's capitalism baby.
And that's why you go with the big guys (and pay a premium for it).
I work for a SaaS company that offers a cloud version as well as a software license. We only support the big 3 because everyone else is just keeping their systems up with chewing gum and duct tape, and it's infuriatingly inconsistent. No way of offering a reasonable SLA or for our support guys to dig into an infra problem. And this includes relatively big players like Ali, Tencent, Yandex, DO or OVH.
In the end, 95% of customers pay less if they choose the cloud version, only if you have 24/7 steady load (and a high one) will it be cheaper to pay for infra, SREs, licenses and support.
Id go so far as to say SaaS in general. Small startups are paying $5000/month to send emails and we've come to the point where inboxes are monopolized and if you don't pay up to a cloud provider your emails end up in spam.
Take this and repeat for everything. Monopolize, ratchet up the costs, profit.
Thinking of those "permissions for Ukraine to strike" being discussed and the reasons Armenia couldn't use Iskander missiles against Azerbaijan in 2020, and Azerbaijan apparently hasn't used Lora missiles after 2020.
I totally agree...the best solution for the specific problem. "Cloud" was the buzzword solution to every problem for a few years and it wasn't great in a lot of cases. High I/O home grown apps to be used from a single campus don't need to be in the cloud. Bulk archive storage doesn't need to be in the cloud, things like lecture recordings from 10+ years.
Completely disagree. This last March, Microsoft changed the storage limit per user on OneDrive for education from 1TB to 100GB, and users either had to delete a ton of files or pay for increased license/space. We ended up standing an on-prem file server back up shortly thereafter because we could not get our users and faculty to delete research data and could not afford to nearly double our cost expenditure. In my experience doing IT budget for years, cloud has meant that you cannot predict your yearly expenditures, Especially if you use your services that are funded in part by venture capital. Let's say you start using some cool research presentation project and suddenly the economy dips and they lose funding, the cost goes way up. Life cycle management has gone completely out the toilets in my experience with cloud products.
It's more jjust a lackof reporting. If Apple came out with something new people would lose their minds. But if some no name Chinese company does it, no one cares.
Melbourne street fashion. Literally asian style pump flip flops with socks half way up your calves. 80s tracksuit baggies. Trying REALLY hard to look like they're not trying. The city is loving it.
I feel like it's hyped just enough. It does have the potential to revolutionize computing but we have no practical applications for it at the current point in its development. There's only so much you can hype something that can't even act as a simple calculator better than a handheld calculator can.
it has the potential to revolutionize some optimization problems that are hard to solve classically. It's going to be practically useless for the average user.
I'd to say that the quality of the hype is completely out of whack. People are expecting the current generation of generative neural networks to do things that they really can't.
The ammount of total excitement is probably actually too low if you group GNNs with AGI, though.
Molten salt sounds like a terrible design for modular, the whole problem is if it loses power it freezes solid, you'd want a huge one with tons of backup imho.
I'd imagine a tiny pebble bed or traveling wave, something fairly inert and safe.
Edit: I guess that's the point, give someone a reactor, if they screw it up it safely freezes dead. Problem solved.
Both the love for Generative AI/LLM is overhyped, but so is the hate for it. They're actually pretty good tools, they won't save the world on their own in their current state.
Thank you! They get trashed on all occasions here in the fediverse and I get the animosity since every corp and their mother now wants to ride the hype train. But I've kinda changed my mind about AI since having been recommended two AI tools that actually cite sources for their answers (Elicit and Perplexity). They're an absolute godsend for the literature search on my Bachelor's Thesis
Here’s how I see it. Gen AI and LLMs are really good for things that I won’t pay money for. It’s undoubtedly impressive tech, but it really deserves to remain as a cool research project rather than an actual functional product.
From what I understand, 5G was first about increased capacity. Increased speed was a secondary point. It optimizes how multiple users can share the same bands, and adds use of higher frequency bands that don't propogate as far. So for very high congestion areas, they can deploy smaller cells and which each can maintain higher speeds per user.
I think the "faster" part was just marketing to get users to buy into the new technology.
I mean I think that was the intent. Something about the implementation needs tuning though.
Not apologizing for carriers, some are really on the edge of lying to consumers, but you have to separate the 2 parts that make 5G different from 4G.
Higher frequencies: means higher throughput but also shorter range (you can literally block that signal with your hand). Only works if your phone supports these higher frequency bands, you have to be in areas where the carrier has deployed cells supporting those, and you have to be close enough.
Increased efficiency: mostly affects carriers, you likely won't notice the difference. Basically means, areas that were congested before with LTE will now see less congestion.
I found most 5G ads infuriating. If you know the tech, you understand whats going on and how they aren't telling the complete story. If you don't know the tech, you'll think, "Yay, higher speeds." Nope...
Passkeys. They'll probably improve eventually but I feel like right now it's a mess.
On Android you are forced to use the default implementation, only in 14 and above can you use password managers for them.
On desktop it's somewhat less messy but you can use the system storage or a password manager extension. Some sites only let you use them for 2FA, some full login, some can't be put in a password manager from my experience and so on.
I am mostly concerned about potentially needing specific Big Tech implementations for them in some way... I don't mind using, say, KeepassXC for it, because it is independent from any account or hardware, as well as easily backupable. But NOT anything tied to a Google or MS account.
Maybe I am misunderstanding something, but Paypal says it restricts what passkeys can be used, so it is apparently possible:
Passkeys are currently available for eligible personal accounts. An eligible Apple or Android device is required to create a passkey.
Other than AI, it’s automation. It’s pretty good when it works but has the same overall intent as AI (in reducing the human labor force), just on a smaller level. At least automation isn’t consistently delivering inaccurate information.
What sort of automation specifically are you referring to? I work in commercial building automation, which is basically tying various systems like fire/burg alarms, access control, energy/lighting management, intercoms, and everything else together using TCP/IP networking, RS-232/485, and dry-contact relay triggers everywhere. For instance, unlocking all doors and stopping elevator access when the fire alarm goes off. Or automatically disarming a burglar alarm and turning on the lights when the first person in the morning scans their badge. In that sense, it works great and has been working for decades.
If you mean robots taking all our jobs, yeah that's about 100 years out.
I was at my company's booth at a career fair earlier this week and it felt like every other student was looking for an internship in "machine learning". When I asked follow up questions about what sort of experience they'd had or projects done or what they wanted to do with it in their career, crickets.
To be fair, 2nd most popular was "CAD" which is also not a job.
Most things to do with Green Energy. Don't get me wrong, I think solar panels or wind turbines are great. I just think that most of the reported figures are technically correct but chosen to give a misleadingly positive impression of the gains.
RiscV is a fundamentally different story then Arm, currently speaking RiscV is not there yet however I have more hope in the future of RiscV then Arm. Both hardware and software side RiscV is not ready however the idea of a fully open source computer still excites me. I understand however that I may be speaking more out of idealism and im certainly biased however I still hope that RiscV overtakes Arm.
Well arm cpu’s get you insane battery life (ie. Macbook M series or new snapdragons). The architecture has not settled in yet but it will take some time
The Arm architecture (Arm_64) which powers Apple and Snapdragon in comparison to AMD_64 (x86_64) which powers Intel and AMD (Intel created x86 and AMD created x86_64)
Electric cars and bigger vehicles. The electricity storage tech is just not there yet. However, I think it's perfectly suitable for personal transportation like scooters and bikes.
For me, the sweet spot is in plug in hybrids, as long as you actually, you know, plug them in. You can cover all your daily commute and grocery getting 100% electric and then if you need to take a longer trip occasionally, you're covered by the gasoline engine. We use like one tank of gas every 4 months on ours.
It wasn’t a very long initial question (only a few sentences), but you somehow missed the only qualifier to the whole thing, “…Besides AI,” within that short intro.