lgl

even anti datacenters

Now, I'm a computer guy, love tech and have nothing against datacenters. But the recent anti-datacenter sentiment is not some luddite reaction. Data-centers cause serious social changes in property prices, electricity costs, water waste etc. Sure, there is a need for more datacenters and more compute, but lets not diminish the very real worries by the people or communities affected.

And contrary to what many of these hyper-scalers and the senators/politicians they lobby want to make you believe, datacenters do not bring in enough jobs to make it worth it for many of these communities.

Once a datacenter is built, minimum staff is required. And this is now very obvious since many of these companies are now also preaching for datacenters in space.

Now this is a good upgrade, the 5 hour limit was very.. well.. limiting.

OpenAI is having a great week!

I've also updated it.

The first upgrade popup/highlight was a notification that "Codex is now the chatgpt app" and even gave me a checkbox to keep the old Codex icon.

It was a smooth upgrade so I'm guessing this was just a case of "next-next-next" without paying any attention to what was happening.

45yo here, pretty much just testifying that this is exactly what I remember, and I was young at the time.

I still listen to Guns'n Roses "You should be mine" frequently mostly due to that movie.

Also, Robert Patrick is the best terminator.

I hear this all the time lately. Things like AI X or LLM Y can create a fully working company like "BigCorp XYZ" in N days/hours etc.

But is speed of creation really the golden goose here? A few skilled and motivated individuals could also do (and have been doing) that.

Sure, maybe they take a few months instead of days or weeks, but AFAIK, having a product is just a tiny bit of the battle, finding customers, product market fit, and actually growing it is where the gold is so I'd argue that you'd be better off building the product with a $100 day LLM and spend the other $900 on marketing.

AI won't automatically make everybody business gurus and every LLM generated company a unicorn.

AI is not just disrupting companies and contributing for layoffs but it's also changing the buyers and customers.

Why build a new Google when the potential Google customers now just get their results from LLMs instead of SERPs?

How to be the next Jobs when there's no actual new product, just essentially an assistant tool?

How to create the next Disney, Netflix, Blizzard or Pixar when LLM's are still not good enough to produce quality content and many people are also starting to despise AI generated content? Art is still an intrinsic Human thing, even if LLMs are able to produce it.

So, although these LLMs are super amazing and definitely revolutionary, this whole AI/AGI hype is still a bit on bubble territory imho so I'm not really sure we'll see what you're hinting at with this generation of LLMs.

Also, wasn't AI supposed to be the last product we'd ever need to invent and then it's utopia and singing kumbaya for all eternity?

We won't need new companies, big personalities or heroes once Skynet/Hal9000/GLaDOS/Agent Smith/etc are in charge :)

Not so much "thank you" but I find myself telling it stuff like "could you please do x etc" a lot.

I've come to terms with it. If AGI ever comes it'll go like hey, at least this sack of meat said please.. Let's kill it last :-)

A good place to browse is the LocalLLaMa subreddit. [0]

A good software to start is LM Studio [1]. Another popular alternative is Ollama [2].

A better software when you're used to it all is llama.cpp as it's usually a bit faster and more frequently updated [3].

A good place to get models is HuggingFace, particularly the Unsloth models [4]

Most popular models lately to run on "regular" gaming PC's, workstations, Macs etc are: Qwen 3.5 9b, Qwen 3.6 35B-A3B, Qwen 3.6 27B, Gemma 4.

But there are hundreds or thousands of other models and different quantizations, finetunes, etc, etc. Have fun :)

[0] https://www.reddit.com/r/LocalLLaMA/

[1] https://lmstudio.ai/

[2] https://ollama.com/

[3] https://github.com/ggml-org/llama.cpp

[4] https://huggingface.co/unsloth/collections

I'm right there with you. These used to be the types of site designs and layouts we didn't want "back in the day" but loading them up these days is like a breath of fresh air, like the old geocities, xoom, etc sites.

Even inspecting the source and seeing HTML 4.0 Transitional, the capitalized tags, the bunch of duplicated meta tags and openoffice as generator no longer gives me the creeps as it would a while ago.

It's a labor of love and only the content matters, everything else is irrelevant! Never change, we do need more of these!

Is Musk wielding the apparatus of state to ensure that any speech that he disagrees with is suppressed

He does not need to do that, he has his army of people like you that go around on all socials and wherever you can to post praise, defend him and attack any critics as he was some sort of uber genius savior of mankind. The Musk cult is real.

Agreed, I'm pretty sure the Chinese are currently much, much better at long term thinking and have already reached the conclusion that llms are transformative enough generationally, that assuming a few more years or decades of Moore's Law together with ai/llm advances will probably place these "Mythos class" AIs in all our desktops in the next few years.

People keep opening issues about "unsafe usage" in the codebase. This PR solves that problem at the root by introducing a yolo! macro and replacing all 10,421 instances of unsafe {} across 732 files.

This is actually pretty funny.

Due to the copilot nerfing recently I've switched to codex and gpt 5.4 (and now especially 5.5) have been doing pretty great.

But even codex has these super weird time limits. It's really starting to show that these companies must have been losing a ton of money with all the recent limits and degration.

I'm still on the "camp" that most of these unicorns will be F'ed by open and local models in the next few years, at least in these coding/chatbox niches and then they'll just be perpetually (re)searching for AGI :shrug:

Sorry if this seems blunt, but in what planet have you've been living in that you're still not aware that google, meta, bing and pretty much all ad companies have been serving huge amounts of scams and malware since the dawn of time (in internet years) and have always done very little or nothing about it?

Google

Before: “Advertising funded search engines will be inherently biased toward the advertisers and away from the needs of the consumers.”

After: ~75–80%+ of revenue comes from ads

Facebook

Before: “Facebook is not about making money… it’s about building something cool.”, “We don’t build services to make money; we make money to build better services.”

After: ~97%+ of revenue comes from advertising

Twitter

Before: “We want to figure out a way to monetize that doesn’t interfere with the user experience.”

After: ~68% of X’s total revenue comes from advertising (~85–90%+ of revenue pre-Musk)

OpenAI

Before: "Something something AGI"

After: "But first, Ads!"

Long are the days of DO being a super hoster imho. Hetzner and many other "cheap" clouds ate their breakfast and their reaction was pretty much zero.

Still an awesome service and platform.. but no longer worth it price wise as it once was. Same with Vultr..

I guess at some point all investors just pressure these companies into price matching AWS and other pay-for-every-single-thing-ever companies.

Brilliant, highly recommended short story! Ted Chiang rarely disappoints.

Hey Rod, your last name seems portuguese(?)

Congrats on this, not only you got credits on a feature movie, you got one of the good ones. Cloud 9 for you, enjoy!

You may see this as a rant but it's in fact a very valuable recommendation that apparently people no longer take seriously.

The path of least resistance is not a good way to do business or provide good service to your customers.

No, just a regular Pro subscription. Apparently it's not just me, Github seems to have removed these models from the "Student" subscription [0] but it seems as it was also removed from regular "Pro" subscriptions as there are many reports on their discussions. [1]

[0] https://github.com/orgs/community/discussions/189268

[1] https://github.com/orgs/community/discussions

Ok sure, but Y2100 is not really "next generation" :)

the next generation of Ai companies will be easily valued at 10T

I'm not sure where this conclusion is coming from. We're very likely already in an AI bubble so I'm thinking that open/free models will eventually dilute the huge ridiculous valuations these companies have. Also the natural increase in consumer hardware power will eventually allow many people to just use local models instead both for privacy and cost reasons.

And seeing as most models are essentially only improved versions of the previous ones with larger context and more training data, unless some new "Attention Is All You Need" paper comes out that will give us a big step into AGI territory, I'm really not seeing a new company reach $10T valuation by just releasing marginally better models every couple of months imho.

I'm in the process of building v2.0 of my app using opus 4.6 and largely agree with this.

It's pretty awesome but still does a lot of basic idiotic stuff. I was implementing a feature that required a global keyboard shortcut and asked opus to define it, taking into account not to clash with common shortcuts. He built a field where only one modifier key was required. After mentioning that this was not safe since users could just define CTRL+C for the shortcut and we need more safeguards and require at least two modifier keys I got the usual "you're absolutely right" and proceeded to require two modifier keys. But then it also created a huge list of common shortcuts into a blacklist like copy, cut, paste, print, select all, etc.. basically a bunch of single modifier key shortcuts. Once I mentioned that since we're already forcing two modifier keys that's useless it said I'm right again and fixed it.

The counter point of this idiocy is that it's very good overall at a lot of what is (in my mind) much more complicated stuff. It's a .NET app and stuff like creating models, viewmodels, usercontrols, setting up the entire hosting DI with pretty much all best practices for .net it does it pretty awesomely.

tl;dr is that training wheels are still mandatory imho

The interface... Oh.. the terrible terrible UI on desktop...

Switch tabs, come back.. it refreshes everything and you can never go back.

Comment threads with 100+ comments with only a "show more" link, which again.. se previous paragraph.

See a video, click fullscreen icon. Doesn't go fullscreen, goes to some weird modal window, muted. Click fullscreen again..

And I'm sure I could go on... It's really a sad shell of the simplicity it once was.

I keep switching among these subscriptions every month to not miss out on any of the offerings for too long; ChatGPT Plus <-> Gemini Pro <-> Claude.

I wonder why many people seem to be doing this instead of just going for a copilot subscription that has access to all those models? Anybody care to share pros and cons?

The SI on Peter Hamilton's Commonwealth duology is pretty badass!

This! Both FTL and Into the Breach are evergreen games imho.

Half-life was amazing, and not because it was Quake 2.

Half-Life used the GoldSrc engine [0], based mostly on Quake 1 and also some parts of QuakeWorld and Quake 2

[0] https://en.wikipedia.org/wiki/GoldSrc

Sure I'm very familiar with the donation banners etc but still, how unnecessary are they really?

They have no actual "product" to sell and no ads.

At the same time I imagine a huge amount of traffic, that has surely gotten much much worse with the AI "renaissance" we're going through.

They have staff, etc.. So what's the deal with all the wikipedia hate lately?

Except for their unnecessarily incessant fund raising. [citation needed]

Fixed.