AlexCoventry

You need to be more specific. OpenAI's commitment to assist the Trump administration with domestic mass surveillance seems to have been largely memory-holed.

Yeah, it's eerie, same with how everyone seems to have forgotten that OpenAI betrayed democracy by committing to work on unsupervised autonomous weapons and domestic mass surveillance.

Is it hitting intermediate milestones with solid pre-written and human-reviewed acceptance tests? If not, sounds like a very risky commitment.

I went to a conference and talked to all the vendors about their products.

According to the article, you can fix this by turning off Gemini Apps Activity in https://myactivity.google.com/product/gemini

I already have this turned off. It's slow, but I use a browser extension to save my Gemini chats locally, when I want to keep them.

> In a subsequent round, it generally can't meaningfully introspect on its prior internal state

It has the K/V cache, no?

It's quite likely that OpenAI is running a significant PR campaign to compensate for the bad rep they earned by stepping in to meet the demands of the Trump administration, after Anthropic refused to assist the administration with mass domestic surveillance and development of lethal autonomous weapons. Presumably OpenAI didn't buy the podcast TBPN just because they like the guys.

https://paulgraham.com/submarine.html

I think there's definitely more scope for ruling out vulnerabilities by implementing simpler designs and architectures.

They have no values of their own, so you have to direct their attention that way.

I don't think "usage" is exactly the metric they're going for, more like "usage in line with our developmental strategy." Transcripts of people using Claude to write code are probably far more valuable to them than transcripts of OpenClaw trying to set up a calendar invite.

I just use Ctrl-g to open the prompt in emacs.

I have a transformer attention mechanism which seems to be more data-efficient than the usual dot product, and I'm trying to write a performant backwards kernel for it.

That's some cynical ignorance.

No, AI has real, immediate, large-scale industrial and scientific applications. Cryptocurrency is a massive disappointment, but AI has genuine potential value.

Yes, it was good. It seems clear that Farrow and his co-author approached it in a methodical, fair-minded way.

https://www.youtube.com/watch?v=wr_sB1Hl0oM

I don't condone violence, but the contract he's signed with the US military is a credible threat to everyone in the US. OpenAI will now certainly be called on to assist in domestic mass surveillance, under threat of the kind of severe penalties Anthropic has faced. So why did he agree to that contract, unless he's will to provide that assistance? So it's gone well beyond conversation, though not to a point where violence is appropriate. Boycotts and hostility are definitely appropriate at this point IMO, though.

Yeah, it's kind of terrifying, how this incident seems to have faded from people's memories.

> The only solution I can come up with is to orient towards sharing the technology with people broadly, and for no one to have the ring. The two obvious ways to do this are individual empowerment and *making sure democratic system stays in control.*

OK! So he's going to renege on the contract he's signed with Hegseth, which effectively commits OpenAI to serving as the IT Department for Trump's secret service?

That's great. It seems like a large body of experts are having real problem with this, so maybe you should publish something about your methods, or start a business...

The hatred is particularly intense on reddit. I lost a couple of accounts there to suspension, just for speaking a civil way about the positive aspects of AI.

I don't condone violence, but it's hardly surprising that people would resort to or support it in this case, considering that by stepping in where Anthropic refused to help the US military, sama essentially agreed that OpenAI will serve as the IT Department for Trump's secret police. Either that, or he's willing for OpenAI to endure a similar punishment when he refuses the inevitable demand to assist with domestic mass surveillance.

> If all of us can go hunting in the woods and yet there is still game to be found, then there's no compelling reason to define and litigate who "owns" those woods.

https://en.wikipedia.org/wiki/Feudalism

Hmm, I still have some nonrefundable API credits with OpenAI. Maybe I should try to use them for my kernel.

FWIW, this talk[1] from NVIDIA/Meta from March claims that coding agents can often write correct implementations of of CUDA kernels, but that they're usually dog slow, like 100x slower than a kernel optimized by a skilled human.

[1] https://www.nvidia.com/en-us/on-demand/session/gtc26-s81653/

Yeah, it's easy to vibecode and review a docker sandbox, too. If you run containers with

   --runtime=runsc
   --cap-drop=ALL
   --security-opt no-new-privileges:true

it's pretty tight. That's how I use coding agents, FWIW.

Do you think they're lying about the vulnerabilities they claim Mythos has found? Seems like a very short-term play, if so.

Try using a coding agent to write an efficient GPU kernel. I guess they might get good at it soon, but they definitely aren't there yet.

Giving up looking is not necessarily the same as not wanting a job.

You can rent it, but it's basically worthless at this stage.

> due to the way round-tripping tool-calls work, splitting code up into multiple files is counter-productive.

Can you expand on that?