xvector

What if they build superintelligence?

Lemonade isn't intelligence. Having 4000 liters of lemonade doesn't give you a competitive leg up over a lemonade stand with 2000 liters.

People can be as anti-AI as they want, they'll still be totally dependent on it. It's not like anyone is pro-Meta, but billions are addicted to IG.

Where do you think almost all of that $67B is going, if not to their infrastructure and inference costs?

If your hyper growth startup does not have more liabilities than assets, it has fucked up.

One day the automated decisionmaking systems will be far better at making decisions than us. That day is not far off. At some point, we will turn to these systems for our most important life-and-death decisions. I don't think that's irrational.

In this era, intending to corrupt the automated decisionmaking systems we rely on, will certainly be sabotage (if done with that intent, as is the case with OP). However, I don't think any sabotage would actually work well in the era of superintelligence, so perhaps the point is moot.

I love how everyone from every political background thinks big tech has it out for them specifically. No, it couldn't be that algorithms are stochastic and random and sometimes ban random people. The tech companies are trying to manipulate everyone, every single side, into some goal only they understand.

Somehow none of the hundreds of thousands of software engineers that have worked there have managed to uncover exactly what this purpose is despite having access to the monorepos and visibility into the recommendation engine pipelines.

On the modern internet, people confuse conspiratorial thinking for well-reasoned thinking. But an idea is not correct just because it is interesting or exciting.

You phrase this as a "gotcha" but I don't think you understand that NGOs and nonprofits are the exact kind of entity that would naturally be involved here.

Algorithmic transparency is a deeply unprofitable and difficult task, with people from all sides screeching at you thinking you're biased against them. Governments obviously shouldn't be involved (unless you're a fan of state-controlled media), and so the only people that will take this on are people that care enough about the impact of social media on society to give up careers in the private sector.

If you care so much, you can always join one of these NGOs! You won't. I guarantee you they are not as biased as you think they are, and they certainly don't hold a "shadow ban this guy" hammer.

Shadowbans have a high base rate. You notice when it happens to your side and take it as a sign of bias. You don't notice when it happens to the other side, because that is not your circle and you don't track them as closely.

I guarantee your political opponents also think Meta is trying to shadow ban then.

Or the public could grow up, learn that algorithms are stochastic, and not let their vote be influenced by the number that pops up when they type in "vote" in the search bar.

If this is influencing your votes your democracy is already lost.

Most current laws around booby traps and sabotage. Intent obviously matters a lot, and under the stated intent this is illegal.

> be to somehow get all world knowledge and tool calls into a audio space, which seems like quite the stretch.

It seems quite easy? They have the natural language bit nailed, they just need to convert the text with it. And it wouldn't be hard to extend architecture this into outputting tool calls.

I don't think you understand how illegal this is. As these models start being used for autonomous decisionmaking in medicine, defense and whatnot, this will be considered industrial sabotage and if it ever harms someone you are so fucked.

I'm not even debating the morality of AI and LLM companies here, this is just fact. If a pharma company is illegally using your water source, you cannot poison it knowing the pharma company will then use this water. Well, not unless you like prison.

If you do this with the intent of LLMs scraping this and this ending up in the corpus used by LLMs, it is indeed very illegal, especially as LLMs become more load bearing for medicine, defense, etc.

It's a bit strange to call it a dead end when this was trained on just a single NVIDIA GB200 NVL72. LLMs couldn't do much back then, either.

It was trained on one NVIDIA GB200 NVL72. This is a tech demo!

Very cool work, I love what you guys are doing!

Why can't Apple just sandbox the damn thing? Run the renderer in a VM, call it a day.

Or just spend a few billion tokens rewriting the whole chain in a memory safe language. This nonsense is inexcusable in 2026

Token burn is insane and it sucks at long horizon tasks.

The Manhattan Project was not democratically confirmed. There are just some things you don't leave to the normies.

Lina Khan had great ideas but ultimately her legacy is eliminating low cost airfare

Citizens are dumb though, the stupidity of the public killed nuclear. Not every decision needs to be democratic. Certainly not core infrastructure decisions.

It's also so easy to astroturf to manipulate public opinion. If I were China, I'd post so much anti-datacenter content everywhere in the West to gain a leg up in the race to superintelligence, because we are dumb enough to let our citizenry influence these decisions.

Imagine if we did the same thing for the nuclear race. ASI is winner-take-all, the first one to ASI can just lock everyone else out.

Most HSMs used by CAs are pretty trash tbh.

Interesting, thank you for the insight!

My job evolved long ago, I'd certainly be worthless at my job if I insisted on keeping the shape of a typical SWE. Automating YOLO merges - we're working on that, I can't wait.

I would say I am now responsible for more big-picture decisions than I was previously, with Claude taking on the role of SWE (or a team of them). Despite this, there is still way more work than people.

No, this is just basic probability.

It doesn't take a genius to realize that you should let anonymous randoms do bio "research" on your API. I'm glad they prevent this unless you are partnered with a university, tbh.

Unironically this is where I converged. Shipping the feature is the most important thing, Opus 5.5 is good enough with correctness that I merge off the title or description alone.

A year from now Fable 6 will de-slopify, so I don't really care about having pretty abstractions. That's simply not my job anymore. Sacrilege, I know.

I don't do this at all. I just YOLO merge after reading the change description. Senior dev, paid well at a FAANG.

I don't really care about the code being unmaintainable. Fable 6 or whatever will take care of it, just like Opus 5.5 takes care of our Opus 4.6 slop. The code working and our feature shipping is way way more important.

They have to scan hundreds of billions of transcripts, would you prefer they just stayed silent for a few months until they scanned it all? Then you would be complaining about them sitting on undisclosed incidents.

There are also legal requirements to disclose within a few days of discovery, at least to the victims. Posting about it on their website lets them front-run any media attention should the incident leak.