He's at the top of the governance

Yes, I hear your point. I guess I take failure of governance to mean that rules and procedures were not followed or never properly put in place--not that the head boss is promoting bad strategy and tactics.

What's the point of using an open source model if you're not self-hosting?

Open source models costs are determined only by electricity usage, as anyone can rent a GPU qnd host them Closed source models cost x10 more just because they can A simple example is Claude Opus, which costs ~1/10 if not less in Claude Code that doesn't have that price multiplier

But Kimi seems so big that renting the necessary number of GPUs is a non trivial exercise.

Exactly! Electricity, hosting, and amortized cost of the GPUs would be the baseline costs.

Open source models can be hosted by provider, in particular plenty of educational institutions host open source models. You get to choose whatever provider you trust. For instance I used DeepSeek R1 a fair bit last year but never on deepseek.com or through its API.

* It's cheaper than proprietary models

* Maybe you don't want to have your conversations used for training. The providers listed on OpenRouter mention whether they do that or not.

The beatings will continue until the stock price goes down.

>beatings continue

>stock goes up

>beatings stop

>stock goes up

For straight up search Google is better, but for AI search I prefer Bing.

> replacing small gasoline generators with silent, emissions-free electricity.

I know more and more electricity these days is produced by green sources, but the statement above which I read often, is never strictly true.

When I had a job at a power plant near where I live, one of the engineers there pointed at the big chimneys and said (I translate:) "what comes out of those chimneys is cleaner than the air around us". They filter out everything except CO₂ with good efficiency and have big, heavy, expensive machinery and engineers onside 24/7 to monitor that it works.

A 5kg motor, optimised for portability, isn't going to be like that.

CO2 is the problem, and it still remains, even in your best case scenario.

The problem? As in, the only one?

A matter of opinion. As I see it the several carcinogenic components of ICE exhausts are problems. Dying of cancer may be a small problem compared to an uninhabitable planet and I can see how one might consider it insignificant in comparison, but in my opinion, cancer is a real problem, including cancer caused by smog.

Where are the emissions generated? Far away from the population of millions.

And much more cleanly/efficiently than a small gasoline-powered generator, regardless of the energy source.

Coal-powered steam turbine is not that more efficient than a portable gas generator so considering coal is more carbon-intensive it's actually about the same or even worse in terms of emissions if you consider coal burning also produces mercury. Now nat gas-powered CCGT - different story. Good news is NYC is mostly powered by the latter and there's zero coal.

> Stephen explained this to me by saying that you cannot make jokes about failure in the States.

This does not ring true to me at all. The overconfident idiot appears all the time in US and UK comedy.

Probably true, but our fintech still gets tons of unsolicited emails from growth equity shops. I don't respond because as you mentioned this sector is out of favor, and as such the multiples are not worth my time.

thats just top of funnel, the instant you respond the terms start to become very very unfriendly

Indeed, cold outreach is to get you excited and then once you're in the weeds, your value is going to get crammed down.

In some sense, LLMs are making me better at critical thinking. e.g. I must first check this answer to see if it's real or hallucinated. How do I verify this answer? Those are good skills.

And perhaps both are overestimating the mean person's ability to detect a hallucinated solution vs a genuine one.

I think hallucination is grossly overstated as a problem at this point, most models will actively search the web and reason about the results. You're much more likely to get the incorrect solution browsing stack overflow than you are asking AI.

Gemini hallucinated a method name in a rust crate then spent several minutes googling the method name + 'rust example' trying to find documentation about the method it made up. Unsurprisingly it didn't find any, and then it just gave up and commented out the entire function and called it done.

Comparing the free tier of Gemini to the latest premium coding models will give you drastically different results.

I'm hoping Porsche's profit woes will lead them to making 911 supply less restrictive--especially in the US. New 911s in US trade above sticker price and have long (not Ferrari long), but long waiting lists.