g42gregory

Are these are friends of Elizabeth Holmes? :-)

While I feel silly to take seriously something printed in The Economist, I would like to mention that people tend to overestimate the short-term impact of any technology and underestimate its long-term impacts. Maybe AI will follow the same route?

Events like this always remind me of President George W. Bush's timeless saying: "Freedom is on the march!"

I tried B-MAD Framework and it was like night and day. Can’t work without it. I’d like to see more frameworks like that.

I have a feeling that China will be OK. Wouldn't these restriction protect the native LLM ecosystem and will give them room and revenues to grow? Strong internal-only competition will still make them competitive.

Good point.

There was a time, when I sincerely thought ACLU was there to protect civil liberties.

These days, I don’t know. Certainly an instrument of a particular party. Plus, owned by various special interest groups beyond that.

As I recall, they were saying the inflation was under control 2019-2024, during which real inflation was around 40% (not annual but aggregate for 4 years).

Same people who say inflation is under control?

My challenge here is that every time I see “The Atlantic” or “The New York Times”, I can’t shake the feeling that it’s got to be a paid advertisement or some influence piece by a special interest group. I am not sure what to make of the articles that appear in those places anymore.

The D here is not exactly defined (or maybe I just missed that).

Does synthetic data count? What about making several more passes through already available data?

I don't understand. The leaders are telling us that, with the advent of AI, we are going to need a smaller workforce. Including in Agriculture, where robotics will play a role in force reduction.

You can't have it both ways: you can't be telling us that AI will force people of of jobs and we will need UBI, while decrying reduction in population.

I respectfully disagree. Google does this by design, as a system designed to provide all legally available information to you. LLMs also do this by design, as a system designed to utilize legal information available to them (coming from the same web as Google's). A criminal can do bad things with a car, a knife, a gun, a kitchen chair and a pencil. Let's not blame the kitchen chair!

So does Google if you ask it.

Don’t forget, we are currently running “Elon is bad” narrative.

Apparently, many people listen.

In both Linux and Windows, websites, in any browser, look “washed out” and not pleasant to look at.

But MacOS always gets it right: in any browser, the websites look “juicy” for the lack of a better word, and pleasant to look at.

Why can’t Linux fix this and render closer to MacOS?

Hopefully, without all the other “value added” stuff.

Shouldn't developer be responsible for every line of code he submits?

He is signing the code with his name when code is pushed.

How is it different from using every other dev tool?

I think of Apple as Linux (POSIX) with a decent UI. It’s not that Apple is good. It’s the alternatives that are so terrible.

On the server with no UI, Linux is the only way to fly.

This is nothing short of brilliant. New styles can encourage to continue learning, while using AI.

I was sort of using this ad hoc before: asking Claude implementation questions after the coding was done. But these styles can take it to next level.

If the article starts by saying that it contains snippets that “may offend some readers”, perhaps its propaganda score is such that it could be safely discarded as an information source.

Nice. I seem to have missed this particular citation. :-)

I don't know where the author gets his information. The GPT-5 is not on par with Claude Opus.

It's still cheaper than Sonnet, but certainly not by 10x. Sonnet, arguably, is still better than GPT-5, although not by much and not in all circumstances.

I read analysis that GPT-5 is better at debugging (super important) and following exact instructions (many people don't quite know what they are asking for anyway, so this could be counter-productive in many cases).

So no, OpenAI did not burn the boats...

(Ok, this is for coding only. I have no information on other uses.)

And you think OpenAI, Anthropic, and Co. will not do exactly the same, when you engage in a wrongthink?

Larry Ellison said it best a few months ago: "AI will keep you on your best behavior"[1].

[1] Meeting with Sam Altman, President Trump, and Masayoshi Son, announcing $500 million infrastructure deal.

Enterprises (depending on the sector, think semi manufacturing) will have no choice for two reasons:

1. Protecting their intellectual property, and

2. Unknown “safety” constraints baked in. Imagine an engineer unable to ran some security tests because LLM thinks it’s “unsafe”. Meanwhile, VP of Sales is on the line with the customer.

This is why it’s so critical to have open source models.

In a year or so, the open source models will become good enough (in both quality and speed) to run locally.

Arguably, OpenAI OSS 120B is already good enough, in both quality and speed, to run on Mac Studio.

Then $10k, amortized over 3 years, will be enough to run code LLMs 24/7.

I hope that’s the future.

I read the section "Hitting a snag: the model builders are also the police now".

It absolutely boggles my mind how anybody thinks that this is Ok?

Unless you are in North Korea, of course.

Interesting. I did not know that.

Boris Cherny was a (main?) creator of Claude Code at Anthropic. He moved over to Cursor about a month ago. I hope Cursor CLI is an Claude Code Agent port to the Cursor. Hopefully, the code quality would be comparable, modulo Cursor's abridged model access. We will know shortly.

I am guessing that price would be almost the same/slightly reduced?

> Mastercard and Visa's assessments were risk-based and entirely apolitical

Any support for this assertion?