halyconWays

This reads like a high schooler discovering that film can manipulate you, but doesn't go any further than that.

"We like to believe that we form our opinions by gathering evidence, weighing facts, and arriving at reasonable conclusions. Be for real. No we do not."

She speaks for herself. The author then doesn't bring up a single fact, piece of evidence, or anything else besides just restating the pedestrian premise a few times. I don't share her inability to study a topic in-depth, come to a conclusion or challenge my previously held beliefs, and then pass judgment while attaching a calibrated degree of certainty to the soundness of my conviction.

It may be a good model, but I'm not reading scrollslop. It annoys me to much when I'm just trying to get information about something.

Everyone was obsessed with classification prior to transformers, then we had 5+ years of everyone (rightfully) obsessed with next-token prediction. What's this sudden resurgence of interest in classifiers? I thought we all agreed that ML tasks generally require something far more advanced than pretrained classifiers. My timeline was also absolutely filled with mentions of Jev, which makes me think it's a successful viral marketing campaign, like langchain. It's now so popular that the dialog is whether or not [frontier company] is poised to catch up to it or not? We already have openjev...anyone can use it. I don't get it, and usually that means it's marketing.

Those sites are also scrollslop. "Slop," as a term, is independent of AI

I call it scrollslop

I never thought I'd see someone else in the wild who remembers Stars! Next we'll be talking about Anacreon and Talisman.

Yes, and when they contain pithy little mic-drop LLM phrases they're intolerable. But I'll still take them over marketing department scrollslop.

I imagine that's why they're pushing them so hard, especially given Google's recent anti-user lockdowns. I have a hard time imagining they'd expend so many resources to protect a tiny percentage of users from having their accounts hacked...there's no way they actually give a shit about that. Their bread and butter is instead protecting the value of their product, which is their captive userbase.

I did LFS once, a long long time ago (days of RedHat 7.3, I'd say, 2002), and it was quite educational! It's a good way to learn the fundamentals of what comprises an actual Linux system

The music that YuE2 is makes is cleaner that Minimax Music 3, but I can't make it generate anything that isn't extremely bland. Songs fail to develop, they always sound like they're stuck in the intro. The composition feels extremely music-school-ey.

>GitHub still remained meritorious enough to merit being bought by one of the biggest companies

You can't just use the word "merit" and claim the CEO's literal statement about meritocracy no longer being one of their virtues means nothing.

"Your honor, my client may have committed grand larceny, but he was certainly GRAND enough to order the GRAND SLAM at Denny's."

>do you even know if they work there anymore?

Yes

>And despite this strong conviction you have, you've provided 0 references (you didn't even cite the original incident, I had to figure out...

So you're unfamiliar with the period of GitHub history you're speaking about as if you have knowledge and authority. Why didn't you ask questions and seek to learn? At what point did you disagree with my theory, before or after you did a cursory search on a single facet of my argument?

>I won't be responding again since I've made my point thoroughly

>Feel free to have the last word

>there hasn't been anything added to the discussion since 3 posts ago

>cheers.

Spoken like a man who feels his argument is defensible and reasoned, clearly, because if this is the manner in which you have debates, you likely always feel that way.

"For extra privacy, you can sign into the Play Store with a Google Account that isn't tied to anything else."

lol. lamo, even.

And yet 2014 was when the company culture completely changed. The co-founder resigned. The company switched to having bosses for everyone (previously they had an informal hierarchy). They moved to having serious HR, PMs, tech leads, 1:1 reporting, etc. They had a brain drain, like firing Zach Holman, a prominent leader. There was that "Dear GitHub" letter about it, when prominent orgs made a public letter about how GitHub was getting shittier. Reliability got worse and worse, steadily. Pre-2014 was called GitHub 1.0. AKA it's institutional rot.

Company: "We're no longer going to focus on merit as a virtue."

Company slowly becomes worse and worse.

Me: "That starting gun surely signaled the start of the decline."

You: "That was a long time ago. Couldn't have been that."

What are second-order consequences?

You don't think that sudden, radical, and politically motivated change in the core leadership of a company doesn't have long-lasting and compounding effects?

Between Gemma 31/26/12/4/2, Deepseek-v4-flash-0731, Qwen 3.8 27B, Qwen 3.8 Flash Next (which I haven't even gotten to run yet!), and now GLM 5.3 Flash, I can't keep up. I love all these open weight models and am continually stunned that it's largely the West fighting for closed, restrictive, anti-user bullshit and China absolutely mogging the likes of OpenAI and Anthropic, with some notable exceptions like Gemma. Still, I shudder to think what the world would look like if we only had closed models. In many ways the stagnation of open source diffusion seems like that: LLMs are just a few months behind frontier, but image gen is like 1.5 years behind.

You mean to tell me it's been 10 hours and there's no unsloth quant? I've been running 0731 and love it.

WearOS watches are trash for me personally as they all require you to run Google Play Services on your phone, which I won't do. The only smartwatch I've had luck with is the Garmin series (Venu, Fenix), and since I really only wear a smartwatch when exercising, it's perfect for me. I find myself happier when I go with an analog watch, and all the notifications I get are never so important that I need to check them the second they come in.

llama.cpp is like the ffmepg of AI, and one of the reasons I so greatly dislike ollama is that the latter completely obfuscates that they're a rebrand of the former. Georgi Gerganov and team did all the hard work; ollama is langchain-like VC-bait with a HF download wrapper.

The only promising social network to me was that browser extension that let you comment on every single page, but it was shut down when people freaked out that it (obviously) sent the URL of every page you visited to their servers. I've wanted something like that for so long. There just isn't enough thread-based discussion happening online. It's all buried in what are essentially darkweb chatrooms (unindexed Discord chats)

You don't need trade routes to invent latrines and garbage cans. European exploitation doesn't create a caste system where it's seen as beneath you to do the most basic of socially hygenic tasks. Obviously there's something more going on

Electricity exists to be used, and we have essentially infinite free energy from nuclear power. Why are we wringing our hands pretending like we're going to run out of electrons? The days of hauling and burning ancient organic material to generate power should be over, and all these guilt-based arguments fallaciously require you to accept that as a premise.

More info on AIPAC endorsements:

Guthrie (4th out of 7): $42k https://www.trackaipac.com/states/kentucky

Pallone (7th out of 10): $241k https://www.trackaipac.com/states/newjersey

"We did multi-shot prompting to try and get these two games into comparable states using these two different models."

"Well obviously you provided better follow-up prompts to the one that came out better."

Also nothing about human-provided plan files and guardrails preclude the one-shot benchmark test. Heavens, I almost said "real coding," but in "real agentic program creation" you'd obviously be doing multi-turn interaction with the agent, but how can you provide a fair test when the model's output n determines your n+1 response?

Long thread but I highly recommend this game. It scratched the itch I had for TA. Well, for everything but that amazing soundtrack. https://www.youtube.com/watch?v=BxAdOQtAFEs&list=RDBxAdOQtAF...

>Click

>Read top link

>It's some garbage about white supremacy

>Read second link

>It's a Hacker News story about Linux

>Same for the next three links

>Read sixth link

>More garbage about white supremacy and how Trump bad

>Close site and don't bother to remember its name

This isn't normal and this isn't how the Internet used to be

So the US government wants Anthropic to require IDs from their users, driving them to over platforms, but won't require this from OpenAI?

>Raspberry Pi are working the issue by releasing new memory variants that are cheaper[2]

Raspberry Pi are working on the issue but letting you spend the same amount of money per GB, for fewer GB.

>How could the maintainers understand their codebase if most of it was not directly written by them?

Counterpoint: I look back at code I wrote a few years ago and just take it on faith that I knew what I was doing at some point. That's still better than never knowing, but it requires faith--faith in a human, vs. faith in an LLM.

How I yearn for an open source alternative to Suno.AI, and something that can create super niche sound effects. This feels like Suno 1.0 levels of quality but maybe it can get there?