willmadden

Watermarking sounds like a good idea, but it's not. Token drift from watermarking will degrade the quality of outputs and could allow clever people to circumvent guardrails.

You should assume all text is AI generated. If you want to "test" someone at school or during an interview, have them write with a pencil and paper.

Haha. Ok.

Then why haven't they caught all of the obvious foreign fraud rings that people on youtube keep uncovering? It's egregious. You would think with their total data access, AI, and Utah datacenter they could at least create a target list for the FBI.

That article is a bit myopic. People and companies don't want to feed all of their ideas, projects, and intellectual property to a morally unscrupulous oligopoly. We learned that lesson the hard way with the last batch of tech monopolies, and the shift towards majority open weight models proves the trend.

Open weight classifiers and open weight LLMs will be burned onto silicon cards in a few years after the models begin to stabilize. They'll be in PCs and laptops. That's going to capture a HUGE chunk of the market.

If you need more horsepower, you'll rent the same silicon safely from AI services cloud providers without handing your data over to Anthropic and OpenAI.

That headline is tabloid tier bs. They haven't established anything close to causation. Do better, science.org.

Ha.

These companies are effectively high-tech plagiarism factories run by CEOs who are competing viciously. Look at their past actions. No, of course you can't!

That's next level masochism.

My concern with this analysis is that the headline "conclusions" depend on a proprietary composite index and assumptions. Where are the underlying data and methodologies for independent evaluation?

I was also surprised by how little attention the analysis gives to China and India. It's like measuring the smoke from a candle while standing next to two forest fires. The real-world impact is to decimate the economies of Europe while industry moves to China and India, who have no meaningful guardrails at all, making the claimed problem worse.

Be careful what you wish for.

The system works because it runs on human ambition. If people in power handed their control to an AI there wouldn't be companies and mergers anymore. It wouldn't make sense. That world could be great or it could be dystopian.

Hello to the new app store and "Nvidia model approval process".

Better headline: Developers prove they have no idea or appreciation for what a CEO does.

Unlikely, but say goodbye to uncensored and abliterated models.

These acquisitions are not healthy. They stifle competition and make it easier for governments to censor open weight models by choking off distribution points.

Were you using the full model or a quantized version, and what harness/configuration were you using?

It sounds like you were using a quant model.

The problem is the FDA's pathetic throughput times. It takes them decades to approve drugs. Blood tests take 1-2 years on average. It's a huge bottleneck. If they could turn these things around faster and work iteratively, it would be different.

As long as we make examples out of the few that exploit, it will be a-ok.

Right, it's paperwork. I get it. Horribly inefficient paperwork.

Google "how was theranos caught" to find out that an investigative journalist uncovered the fraud, and the FDA missed it.

Their doctor should be qualified to make that judgment without a government bureaucracy evaluating the efficacy of the test on their behalf, unless you want to completely destroy the rate of medical breakthroughs. If there's no potential harm from the procedure itself, the government shouldn't be involved at all.

Imagine if every new AI model had to be "cleared" by a government regulator. The role of government in healthcare evaluation needs to be simplified and streamlined. The liability needs to be shifted to doctor, and if it's a case of charlatanism, that liability should include a lengthy prison sentence.

Why is the FDA "clearing" something that is completely innocuous like a blood test?

What about it?

You missed the Terra Invicta "you think it's easy, but you'll be dead in 40 hours because you are underperforming" curve.

No, you are putting words in my mouth. I would gladly pay more taxes for representation that had me and my peoples' interests in mind. I'm not a libertarian idiot.

You didn't read my post objectively. My point is that the current governments, with the exception of some local governments, do not represent or serve their constituencies; in fact, they actively subvert them. I'm assuming your European, that's especially true for you.

You are claiming infrastructure can only be maintained by a government? That's a ridiculous assertion. Most of the work is outsourced to contractors. They are acting as mediocre project managers, not the masters of inscrutable infrastructure technology.

The state has a monopoly on certain types of infrastructure that it ultimately sustains through the threat of repossession. It uses that monopoly to racketeer property owners under the threat of repossessing their property.

If what you are claiming is true, then why does Malta have higher property values than countries with property tax?

https://www.maltasothebysrealty.com/property-for-sale/in/sli...

I would agree to a property tax if the voting systems worked fairly at a sufficiently granular level, but they don't. They are farces maintained by criminal organizations beyond a certain scale.

You don't really own it because of the property tax. Unless you live in Malta or a handful of other countries you are effectively renting from the state.

Yes, then we can build a company store where the slav... er, "employees" can shop and discuss how to be more productive on their projects!

Thank you.

Anthropic and OpenAI have proven themselves to be aspiring monopolists through their positions and actions. Why would you want to share your research with them?