sleight42

What percentage of whistleblowers are punished, directly or indirectly, for their ethical behavior? What percentage come out of the experience worse? Worse finances? Worse health?

The incentives aren't there to behave as an ethical human being in business beyond getting caught breaking the law (for a narrow definition of ethics).

I say this as someone who fought for the smaller fry multiple times (one sexual harassment victim, one employee at risk of losing their visa, and multiple teams getting done to) in the workplace and came out of it with autistic burnout (I didn't know then about my autism) from the emotional labor leading to unemployment.

The incentives aren't there. I wish they were.

̶D̶o̶n̶'̶t̶ be evil

Please modify to include multiples of each as well. Most serious people aren't going to use a single PCIe card if they can help it.

HN users excel at believing their competence in tech applies to competence in every other field. We're an arrogant delusional lot in that way.

AI works. It's that Apple is serving up shit models because they're giving it away for free. Can't expect much from a 9b model. Now the 172b I'm running at home works great.

It's not "likely". It's real. Junkets, for starters.

Look no further than Supreme Court Justice Clarence Thomas for a plethora of modes of corruption used in the American system.

American here. This is bullshit.

How often to the non-moneyed interests win the legislation game? Money gets you face time and attention while paying for politicians' political campaigns so they can stay in power.

It's corruption.

Yes, all citizens deserve access. And they have it, Except it's Animal Farm: all pigs are equal except some pigs (the rich, the "Super PACs", and the corporations) are more equal than others. Because they spend money for influence.

Corruption.

Mosquitos? I think of the massive bloodsucking politicians and billionaires...

I'm on the other side of the political spectrum here AND YET I entirely wholeheartedly support what the person above wrote: use self-hosted LLMs, abliterated or otherwise. M

For many, AI chats are damn close to extensions of our minds: diaries. Those are supposed to be private.

Was what that woman wrote a credible threat? Was she blowing off steam? Without knowing her deeply, how can you tell?

Maintain your privacy. Prevent thought policing.

Just don't try concurrent requests. It's not fun. It does not seem to process in parallel at all. Instead, I'd see it swap contexts out frequently. Inefficient AF. Makes sense if you consider that each turn between two trying-to-run-in-parallel requests would activate different experts, requiring different parts of the model to be loaded to service different requests.

Maybe there's some whackadoodle way to only use a portion of the VRAM for experts for one request and another portion for experts for the other such that requests could actually run parallel instead of concurrently?

EDIT: I'm wrong! It already can do this with the parallel argument.

However, a few days ago, the dev(s now?) added swapping contexts to and from RAM.

Also, from my above, I don't see wh

Anecdotally, I've found that 27B at 4bit hallucinates a lot more than QFN at 3_xxs, particularly for less technical reasoning.

I've tried to use both for online comparison shopping. QFN not only seemed less delusional but also made useful observations and problem solved ways around many different website access issues.

Moments of brilliance yet unpredictable versus reasonably good (lets be fair: 27b models are pretty damn good now) and predictable.

I'm an old school engineer. I like predictable boring technology that consistently gets the job done over hotness. A 27b model at high quant is "hot" enough.

Speaking in terms of teams: I hate working with hotshots. They ruin teams. And they're usually inconsistent and bad for morale.

I don't do that. I use my 2020 top end gaming PC with its 3090. I just ordered 128GB RAM for it. I'll use a single-chat/slot runtime like Strata or Freetoken. And then I'll be able to run either a blazing fast 8-bit Qwen 3.8 27b or a 4 or 5 bit Qwen 3.8 Flash Next. That's good enough for me for development.

And then I'll have my current gaming rig with less RAM but better CPU and GPU run a reasonable smart tool agent for handling Home Assistant Voice Assist. Downside there: when I'm gaming, no voice assist. That may piss the wife off. We'll see.

At least, this way, I don't have to worry about god damn usage limits. I can knock myself out.

iPad app and then we'll talk. ;-)

But, seriously, I'll fire this up!

Right. Except not.

There are plenty of self-hostable models. Hell, there are now model-specific runtimes that make it easy to run big models on consumer hardware. Strata, just a couple days ago released, lets me run Qwen 3.8 Flash Next at home on a single 3090 at 90 tok/s.

So, yeah, I'll be passing on these "labs" walled gardens,

It's always about increasing monthly active users, locking them in, and then enshittifying to squeeze out profit.

Thanks. I'll stick with self-hosting.

Zuckerberg needs to fix Zuckerberg first. Meta is going to remain as it is, otherwise.

The irony of flagging this post should be obvious.

If you're not already, I strongly recommend using [algolia](https://hn.algolia.com/?query=&sort=byPopularity&prefix&page...]) to view HN. That way you can still find content that is ordinarily rudely shoved out of view.

Another irony: using algolia, I've come to see most of HN users as refreshingly liberal. So it's the site itself that buries opposing views.

JFC this was clickbait. Even if this wasn't the intent and even if the dude has atypical sexual practices, if it's between consenting adults then how is it anyone else's business?

I cant remember the last time I flagged a post. Well, here I am.

For the life of me, I can't understand why your comment is being downvoted or flagged or whatever makes it go gray on HN.

I'm guessing it's people knee-jerking that you're being political?

I can't understand the people who don't see it.

The data centers strain the power grids then electricity costs go up for everyone else. This is de facto a regressive tax because everyone needs electricity and the poor pay proportionately more of their income for the increased cost.

Live in San Francisco? Probably not now unless you're rich because of the skyrocketing cost of living due to Tech and AI money. Another de facto regressive tax, driving away other people.

Environmental damage? The poor are the most impacted and the least able to absorb the costs. Do they have the property or renter's insurance to protect them from these disasters? Another de facto regressive tax.

Need a new phone? Or a computer? Same problem.

Want to dabble in AI? You're not going to get too far on $20/month. It's mostly a wealthy person's game.

Or there are the statistics that the vast majority of successful founders from up upper middle class families or wealthier.

Wealth centralization is what our economic system does. The purpose of a system is what it does. If it wasn't, the system would have been changed.

I'd distill that to: Larry Ellison going bankrupt would be the universe rendering a rare justice.

Real question: have you asked it to do basic arithmetic? Does it work for you?

I do NOT understand the bullishness on Siri. I've been running the dev builds for month. The thing I hear the most often is, "I'm sorry. Something went wrong."

Seriously? No one said "the first step in Roko's Basilisk"???

You can always buy the $2000 edition. Or give Apple the finger. I'm leaning toward the latter.

13 Pro here. Just upgraded my battery. May go another year or two at this rate.

Always said that I didn't see a reason to upgrade once I had 5G UW

A $2000 phone. Read the room, Apple. Fuck off.

If I read "load-bearing", "pin", "wedge", "rung", or "it's worse than what I'd previously stated" one more fucking time from Opus...

I've had Claude review output from Qwen3.8-27b. Claude negged it! Said it was hallucinating!

27b may be small but it seems competent most of the time.

Cynically, I wonder if it's a way to increase token spend from repeated requests to speak like fucking human.

The number of times my response has been "Plain English"...

I started using "debuzz", a skill that runs Claude output through antigravity. Works. But makes everything even slower.

Anthropic needs to get their shit together.