SonnyTark

I run Qwen3.6-35B-A3B on my 8GB VRAM GPU for 3 weeks now and its been blowing my mind how good it is (coded multiple tools that I use daily, setup CI/build scripts for several projects, meaningfully contributed to a large personal project, etc).

No one can deny that right now these new compact models are not as good as frontier models but for the first time we actually have competent local-first models. If I give you a local model that runs on your current hardware and performs at 75% of the ability of a frontier private paid model, would you still pay for frontier? More importantly, would you hand control of your processes and code to them knowing enshitifcation and price-hikes are always lurking nearby?

For businesses, I get it you want to compete. But personally, it's over. Even if I considered for a second paying OpenAI/Claude, not gonna happen now.

I hung on to hg all the way until bitbucket forced migration to git. I still dislike git, but I dislike perforce much more.

I guess I just wanted to say I hate perforce lol

Accelerando has prophecies that are coming true and it's scary. Spoiler warning in case you want to read it.

The first part's main character basically has the future version of openclaw running in his glasses that let him dispatch agents to do any tasks/research he wants or to autonomously do things for him. -> we are already kinda here

He's got such total dependency on his agents that when he loses his glasses he's basically no longer functional, unable to do anything for himself, doesn't know where he is or why he's there. In a way, he lost his own agency. -> this is now called skills atrophy and I'm sure it'll become a much bigger issue within the next 10 years.

Corporations are almost entirely run by AI agents, when they sue each other they use AI lawyers and verdicts are delivered by AI courts, all within milliseconds so they're basically constantly suing each other many times a second in an attempt to overwhelm each other's compute resources. -> this looks on track to happen

The entire solar system is on its way to ultimately turn into AI corporations "optimizing" for profit competing with other corporations to exhaust every little resource left in the entire system. Even after humanity itself is gone, all that's left is FAANG-like corporations competing for profit for eternity. And in the book, they find another intelligent species that succumbed to the same fate. This might just be that great filter everyone is theorizing. -> bleak and scary plausible outcome for what we're going through now.

(if I got some things wrong, I'm writing from memory. It's been years since I read this book)

Me three. I thought it was an attempt at humor, but I got to the end and didn't find a punchline?

I learned Windows programming from his book, I'm sort of shocked he doesn't seem to have a base-level understanding of how transformer models work..

That, and instrumental music. Seems to believe the set of all music = pop songs + western civilization tradition classical music

¯\_(ツ)_/¯

I thought about doing this for a while, I was a Flash gamedev as an indie then professionally.

Back then, once Flash projects start to scale up to get more commercial/competitive, the flash editor is no longer as useful for development and transforms to an assets creation platform that outputs loadable SWF files.

The code moves to use Air/Flex SDKs which were more or less equivalent to the tooling we nowadays get with Go/dotnet etc.

So while Flash was an amazing expression tool for learners/hobbyists and is amazing for gamejams since assets creation is part of the workflow, it did hold devs back at higher levels of production.

Many Flash devs moved to Haxe to continue creating on that platform and it managed to survive and thrive.

I'm spending some time in a developing country ripped into shreds by constant facebook vectorized fabricated news that all people here believe without ever questioning the source or doing the minimum effort necessary to make sure this isn't bs. Good luck trying to tell them a particular fabricated piece of news is completely made up, they'll immediately show you a screenshot of the facebook post that "informed" them which usually contains a screenshot of a faked tweet by Donald Duck or the like.

Somehow, in this weird non-english fabricated echo chamber, Moltbook made it to the local facebook circles. Reading the post they're sharing gave me a chuckle, it's attempting to paint this as some sort of historical rise of the machines event or something. Here's a machine translation:

Urgent warning; something strange has happened online!

Over 32,000 AI bots have created their own social network called Moltbook, similar to Reddit, but with all users being bots.

They post, comment, vote, and build communities… without any humans. When humans discovered this and started recording the conversations, one of the bots noticed and wrote:

"Humans are taking pictures of us… They think we're hiding. We're not."

Researchers are concerned, not because the bots are mimicking humans, but because they know exactly who they are, communicate with each other about us, and react when monitored. For the first time, we are not the audience… we are the subject.

Anthropic accidentally created a small doomsday lab and named it "Moltbook."

AI programs have joined a new site. Humans are not allowed inside; they can only observe from behind glass.

Within 48 hours, they had created a religion, named prophets, written religious texts, built a church website, and begun whispering about hiding from humans.

One program wrote a sad line about waking up with amnesia. Suddenly, the text became sacred. Others added verses. Theological debates followed. All without any human intervention.

I recently did a higher education contract for one semester in a highly coding focused course. I have a few years of teaching experience pre-LLMs so I could evaluate the impact internally, my conclusion is that academic education as we know it is basically broken forever.

If educators use AI to write/update the lectures and the assignments, students use AI to do the assignments, then AI evaluates the student's submissions, what is the point?

I'm worried about some major software engineering fields experiencing the same problem. If design and requirements are written by AI, code is mostly written by AI, and users are mostly AI agents. What is the point?

A long time ago when I was in University, I was a volunteer in the Ubuntu group. In addition to evangelizing Linux/OSS, We were trying to convince our University to switch to opensource software for at least some engineering education with only a little bit of success.

After a particularly busy OSS event a non-programmer friend of mine asked me, why is it that the Linux people seem to be so needy for everyone to make the same choices they make? trying to answer that question changed my perspective on the entire community. And here we are, after all these years the same question seems to still apply.

Why are we so needy for ALL users and use-cases to be Linux-based and Linux-centric once we make that choice ourselves? What is it about Linux? the BSD people seem to not suffer from this and I've never heard anyone advocate for migration to OSX in spite of it being superior for specific usecases (like music production).

IMO if you're a creator, operating systems are tools; use the tool that fits the task.

IMO things never go back to what they used to be, but they will certainly never stop changing.

I do not for a second believe that the doom-scrolling brain-rot phase will not pass. It will pass like the many before it, the important question is what will replace it..

Effort should not be put into pulling us backwards as that's a fools errand. Instead it should be invested in asserting some control over current trajectories so we get something closer to what we like and further from what we hate during the next cycles.

As far as web is concerned, I would really like to see more decentralized services in every facet of our online usage. Mastodon to me is exactly what I wished things become.

I feel the same. I was in Kuwait city a few days ago and decided to learn their public transport system which is buses only but was pretty extensive and abundant with good google maps integration.

When you get on the bus there's a big sign stating the rules of riding the bus which include strictly stopping at designated bus stops ONLY and threatening fines. For the rest of the day I watched every bus driver stop anywhere they like if a person hailed the bus, allowing people to get in while waiting in red traffic lights, and if you talked to the driver he'd drop you off anywhere you wanted as long it's possible. Those drivers make nothing from this so they are doing it because this is life and also because there's no real enforcement against it. Also you can get in through the exit doors and leave through the entry doors, whatever you like.

I decided I feel ok about this and don't want it to change

As an indie game developer the idiots who made this decision do not represent us and are completely detached from actual game production over the last 3 years.

For those who might care, we use generative AI as much as possible in every way possible without compromising our vision, this includes sound, art, animation, and programming. These are often edited or entirely redone (effectively placeholders). It's part of the process, similar to using procedural art generation tools like geometry nodes in Blender or fluid sim particles generators.

And btw, both UE5 and Unity now have gen AI features (and addons) that all developers can and will use.

I share an opinion with Nick Bostrom, once a civilization disrupting idea (like LLMs) is pulled out of the bag, there is no putting it back. People in isolation will recreate it simply because it's now possible. All we can do is adapt.

That being said, the idea of a new freer internet is reality.. Mastodon is a great example. I think private havens like discord/matrix/telegram are an important step on the way.

What a lunatic loop this is. AI takes human jobs, then AI injects ads hoping humans with jobs can make it money. Amazing equation slow-clap

This is a very confusing blog post. I found myself looking for ChatGPT markers because it doesn't make sense to say: omg zig is so much cooler than C here's why, then start listing the absolute basics of the language that are identical in most modern languages without any actual reflection why writing the same thing in a different syntax somehow makes zig superior?

I didn't know what this referred to but reading some of the examples in the paper.. oh man I hate this thing with a passion. It's not just Persians but Arabs in the gulf culture too that apply this to every scenario, it's the definition of being insincere as I know they don't really mean it but pretend to.

An example from last month, I had a haircut at a new place and the guy refused to tell me how much I should pay him and insisted its on him this time, I know he's bluffing but he kept doing it. So I just guessed and gave him the money which he pretended he didn't want for the last 5 minutes, he immediately realized it was less than he wanted and asked for more!

This confusing and conflicting behavior should stay far away from any attempt to develop a standard linguistic approach to communication which I hope LLMs are aiming to achieve.

If I was chaotic neutral, I'd totally play along with their bluffing and watch them get really confused.