michaelcampbell

That's String Theory providing "answers", no?

It isn't. $COMPANY I've worked for use commit counts as a metric, and you can bet all the money in your pockets they've skyrocketed with no change to actual output after they did.

LLM's make it even easier; "Commit all the outstanding code in as many commits as you can, as long as the tests pass after each one". (Sometimes that second clause is ommitted, too.)

Not to diminish your lived experience, but this is sort of a management issue, no? There should be some expectations set and/or gates that need to be opened during PR about commit expectations. Some teams like a lot of commits to be a (messy) reflection of actual reality/history, others want the well-curated "story" with a bunch of squashes and rebases.

`jq` and `ffmpeg` were two of my very early uses of LLM's for daily ease.

This feels like a Gwern blog, or something he'd be interested in.

multiple things can be bad at the same time.

I made no such assertion. Only that businesses do things in the business's interest more frequently than databreaches.

My point is I don't think one bit of this is accidental.

Naming all your products with X because it uses some fashion of X is certainly a choice.

I think Satya has lost the thread, even in a CEO context.

> I mourn the loss of working on intellectually stimulating programming problems, but that’s a part of my job that’s fading. I need to decide if the remaining work - understanding requirements, managing teams, what have you - is still enjoyable enough to continue.

I am in the same boat, but close enough to retirement that I'm less "scared" about it. For me I'm moving up the chain; not people management, but devoting a lot more of my time up the abstraction continuum. Looking a lot more at overall designs and code quality and managing specs and inputs and requirements.

I wrote some design docs past few days for a big project the team is embarking on. We never had that before, at least not in the level of detail (per time quantum) that I was able to produce. Used 2 models from 2 companies - one to write, one to review, and bounce between them until the 3 of us agree.

Honestly it didn't take any less time than I would have done it alone, but the level of detail was better, and covered more edge cases. Calling it a "win" right now. I still enjoy it, as most of the code I/we was/are writing is mostly fancy CRUD anyway, and doesn't have huge scaling problems to solve (and too few devs I feel are being honest about their work, here).

> > BrowserStack routinely sell or give away their users' data.

> Or the simpler answer, their db/email list has been compromised.

I find the first option far simpler.

> not realizing the privacy implications.

If only.

I seem to recall he had a cadre of people actually writing the browser, too. (Also Indian, as I remember, but it's been awhile.)

It could be a faulty memory, but I've never thought much of the guy after I heard that, and he continues to not impress.

They have no marketplace to religiously defend for it...yet.

Fun. Interestingly the one that "won" for me isn't the one I normally use, which was one of the candidates, but I have used it a lot in the past.

Same, and constructing at least drafts of huge documents that I can iteratively fine-tune that have (at least last week) saved me 10's of hours.

And based on reality (code) rather than my feelz of what I vaguely remember the code to have been doing in some long past.

My take is new capabilities will consume any price reductions, making them moot. At least in the medium term.

A RAM price drop due to some magic efficiencies assumes everything else doesn't change, which I doubt anyone honestly thinks will be the case.

I didn't have any "top of page" navigations that I didn't control going on.

So no evidence then. Kind of like Lex touting his bona fides as a professor.

Do you have any links to documentation of this? Andreesen has a definite bias as well, so I'm not about to just accept his say-so in a fit of Appeal to Authority.

(eg: "Cite?")

> We've been securing our systems in all ways possible for decades and then one day just said: oh hello unpredictable, unreliable, Turing-complete software that can exfiltrate and corrupt data in infinite unknown ways -- here's the keys, go wild.

These are generally (but not always) 2 different sets of people.

> Our uptime has a '9' in it! -- Anthropic

Dumb enough to show a splash screen that might last anywhere from an infuriating time to nigh infinity, I guess.

That it's designed for a thing and becoming the go-to choice for that thing can be far apart indeed.

> Swift could have easily dethroned Python.

Just IMO, but... no. To me a "could have easily" requires n-1 things to have happened, and 1 thing not happening. Like, we "could have easily" had a nuclear exchange with the USSR, were it not for the ONE Russian guy who decided to wait for more evidence. https://en.wikipedia.org/wiki/1983_Soviet_nuclear_false_alar...

But even in '15-'17, there were too many people doing too many things with Python (the big shift to data orientation started in the mid/late 90's which paved the way to ML and massive python usage) by then.

The 'n' was large, and not nearly of the 'n' things were in Swift's favor then.

Again, IMO.

I want to accuse you of using an LLM to write this with the temperature set to some absurdly high value, because on its face it sounds ridiculous.

And yet, here we are.

Doesn't mean 10% of any crashes either; the OP narrowed it down to roughly 1 in 20, then pulled a factor of 2 out of, well, thin air to get 10%.

At least 10 years ago, some IBM tools I was forced to use worked faster in a linux VM on Windows (through VirtualBox) than on Windows.

> I think that something in Google engine is profoundly broken

Optimizing for ad revenue is a good start.

"This is AI" is the new "This is 'shopped, I can tell by the pixels."