This isn’t news. We’ve known this for many, many years. It’s one of the reasons why many companies didn’t bother using LLM’s in the first place, that paired with the sheer amount of hallucinations you’ll get that’ll often utterly destroy a company’s reputation (lol Google).
With that said, for commercial services that use LLM’s, it’s absolutely not true. The models won’t reason, but many will have separate expert agents or API endpoints that it will be told to use to disambiguate or better understand what is being asked, what context is needed, etc.
It’s kinda funny, because many AI bros rave about how LLM’s are getting super powerful, when in reality the real improvements we’re seeing is in smaller models that teach a LLM about things like Personas, where to seek expert opinion, what a user “might” mean if they misspell something or ask for something out of context, etc. The LLM’s themselves are only slightly getting better, but the thing that preceded them is propping them up to make them better
IMO, LLM’s are what they are, a good way to spit information out fast. They’re an orchestration mechanism at best. When you think about them this way, every improvement we see tends to make a lot of sense. The article is kinda true, but not in the way they want it to be.
The man spends his time mostly on Twitter, driving it into the ground to simp for Trump. He’s taken obscene amounts of money from it, while failing to offer the simplest of products from its roadmap a decade ago - an affordable electric car.
If they don’t pivot in the next few years, they’ll run the risk of Tesla being leapfrogged to market more than it already has, with an inferior product to practically everyone that enters the EV market.
I still fully maintain that Musk will push Tesla towards making a petrol-powered car within the next few years, so IMO the board needs to get rid before that becomes a thing.
One of our engineers: Unbelievably helpful, all the time, to several dozen people. I have no idea how you find the time, but for someone in your position to respond to messages to request for any help is an achievement in itself.
One of our managers: I’ve worked with you for 3 years, and I have absolutely no idea what to do day-to-day. I also have no idea how you’re still in a job. I’m not the only person either.
I used Reddit for around 15 years, and had never seen the above. Sure, it was toxic in other ways, but Lemmy in 2024 is basically 2004 Slashdot - just replace Bush with Trump, and it’s basically the same shit about how Microsoft is evil, anyone that doesn’t use Linux is a moron, and arguing about Star Trek.
Not quite. .NET is owned by the .NET Foundation, and while it’s heavily influenced by Microsoft, it’s an independent entity. C# is owned by Microsoft, but frankly they’ve put together what was even then far more advanced than anything Java could do even now.
To be blunt, back in the 2000’s it was this exact mentality that pushed me towards C#. Instead of people bitching and starting holy wars about Java, Ruby, and other languages, the .NET community just quietly got on with things and built some fantastic tooling. Furthermore, it was one of the communities that helped me go from hapless junior to someone able to give technical talks on what I had learned, or even speak to giants in the industry like Jon Skeet.
Easy, because it’s often home to some extreme and often bizarre opinions. It often makes Reddit seem civil and intellectual.
In the last month alone, I’ve seen:
Downvotes for suggesting that if you were to ask Big Tech to split, they’ll probably just split in a way that keeps them aligned, while cutting unprofitable chunks out.
Downvotes for suggesting that TikTok shouldn’t be banned for being “stupid”, given that most social media is stupid, and people often said that about the shit that many of us grew up with.
Just endless nonsense about how you cannot print with Windows, like at all, while Linux (never stating a distro) “just works”. It’s clear that so many people in the tech community on lemmy.world haven’t used Windows for years.
Crap about LLM’s and how providers like OpenAI are “dumb”, despite the fact that many use grounding and expert systems to guide towards correctness (literally what I work on in my job).
Being so US-centric AND contrarian that you regularly see posts around how “Europe is so much better for this”, when Europe is a fucking continent with separate countries and their own laws/customs.
Posts about how we can get “normies” to use Lemmy
None of the above would happen on Reddit. They’re all signs of communities that are detached from reality, so much so that on Mastodon there are several posts from people that have called Lemmy (basically meaning the “main” instance) out as being toxic and unfriendly compared to other fediverse offerings.
I actually used Drupal a year ago, so it’s definitely still around! Joomla isn’t a name I’ve heard for a while though. To be fair, I mostly work in AI now, so I’m removed from the web dev world also.
I think flat file and API based CMS’s have become more popular now, especially with many people questioning why so many CMS’s were built on relational data stores for largely non-relational data. For many, the ability to drop a CMS in and have it “just work” is why some of the newer ones are growing in popularity.
It’s not early 2000’s Slashdot. .NET and C# have been solid choices for software development for years, and Umbraco in particular is open source and probably the most welcoming CMS I’ve known when it comes to contributions.
I am somewhat biased, as an employee of a big tech company - but I am okay with them moving into different industries as long as they don’t undercut while also providing just the worst employment experience of all time. It sucks to see nice startups from passionate people get steamrolled by a 100 person org full of people fearing for their job while some exec rides the coattails of their boss.
I’d be more supportive of big tech if they were nice places to work, but many of them simply aren’t. They have “prestige” (whatever the fuck that means), but some of them are full of some of the most broken, beaten-down people you’ll ever meet.
If you want a standard CMS, you can’t really go wrong with Umbraco. Some people are turned off by .NET, but for developer experience alone it’s the best I’ve ever worked with.
There are many good choices, if you’re looking for something more lightweight. Kirby, IndieKit, Concrete5, even Ghost are all solid. I also remember hearing about ClassicPress a while back, that was a fork of WP made during some technical and business decisions that some in the community didn’t agree with - never used it though, and it’s a fork of a time when the WP codebase was a joke.
As someone that made enough money to make a freelance career from moving people off of awful WordPress sites, WP’s reputation has been in the toilet for a decade, easily. The CMS market has been strong for a long time, and there are countless better options out there.
With the push towards API backends and static sites, WP should have died years ago. I still cannot believe it’s so popular.
No shit, good fucking luck getting a business to purposely neuter itself.
Any reduction in operations or separating into new businesses would almost certainly be an effort to trim expenses/fat, and not a realistic effort into creating multiple viable businesses.
With that said, I’d definitely cut Cloud. They’re a distant and expensive third to AWS and Azure, and it probably doesn’t make the kind of money that other arms will make.
While I do think that many of these companies need regulation, I think it would be very easy for many of them to cut off a finger or two to save the body, especially when you factor in that many departments of these companies either operate at a loss, or are in positions where they are losing market share.
For Google, losing Chrome would do very little for them. Fill the board with several execs, and it’ll be Google-aligned for the next decade or so. They could also kill off Music, Docs, Fit, Pay, Keep, almost a dozen products that could either be killed or spun off into separate businesses. The same goes for Microsoft, Amazon, Meta, countless businesses that have a finger in a lot of pies.
EDIT: This is why people on the fediverse don’t like Lemmy…
“sigh”
(Preface: I work in AI)
This isn’t news. We’ve known this for many, many years. It’s one of the reasons why many companies didn’t bother using LLM’s in the first place, that paired with the sheer amount of hallucinations you’ll get that’ll often utterly destroy a company’s reputation (lol Google).
With that said, for commercial services that use LLM’s, it’s absolutely not true. The models won’t reason, but many will have separate expert agents or API endpoints that it will be told to use to disambiguate or better understand what is being asked, what context is needed, etc.
It’s kinda funny, because many AI bros rave about how LLM’s are getting super powerful, when in reality the real improvements we’re seeing is in smaller models that teach a LLM about things like Personas, where to seek expert opinion, what a user “might” mean if they misspell something or ask for something out of context, etc. The LLM’s themselves are only slightly getting better, but the thing that preceded them is propping them up to make them better
IMO, LLM’s are what they are, a good way to spit information out fast. They’re an orchestration mechanism at best. When you think about them this way, every improvement we see tends to make a lot of sense. The article is kinda true, but not in the way they want it to be.