shermantanktop

Easy way to start: ask AI to make the codebase more robust. Modern models are quite good at finding areas of fragility.

So many of these complaints suggest that people are accepting that the first attempt by a model is the best it can possibly do. Yes, it takes effort to guide and correct the model's output, but not that much effort.

You may not want that to be your job, which I understand. And AI isn't great at everything, and has gaps, etc. But if you are going to give AI a fair shot, don't hamstring it by using it poorly and then claiming your bad results are representative.

The author, Felix Rosenthal, was a family friend. He had an incredible life, like something out of "The Third Man":

Born on January 12, 1917 in Munich, Felix emigrated from Germany in 1933, shortly after Hitler came to power, and settled with part of his family in his mother's native Florence; after obtaining his Italian high-school degree ("licenza liceale") he went on to study engineering at the Istituto Politecnico in Milan, but when the Fascists' anti-semitic laws of 1938 prohibited Jews from enrolling in public schools and universities, he settled in Paris, again with part of his family. At the outbreak of WWII in September 1939, Felix was in Switzerland and a return to France was impossible; in 1940 he emigrated to Chile before finally settling in Berkeley to join his brother Bernard in 1941, and enrolling in the engineering school of the University of California. From 1943 to 1946 he served in the US Army and became one of "The Ritchie Boys" - GIs of German origin who spoke German natively and were trained in Military Intelligence at Camp Ritchie, Maryland. He was discharged with the rank of Master Sergeant and returned to the University of California to study architecture; after having obtained his MA he became assistant to Eric Mendelsohn, the renowned architect who had recently joined Berkeley's Architecture Department.

https://www.legacy.com/us/obituaries/legacyremembers/felix-r...

We had the original book when I was a kid. My wife gave me a copy (off of AbeBooks or somewhere) a few years ago and I was a bit overcome with emotion at seeing it again.

> try "choosing" to be attracted to an 80 year old woman

Not a great example. Elderly women may not be your thing, but nursing homes are apparently hotbeds of STDs. There's someone out there for everyone.

Oops, yes.

US coverage of foreign elections can't indulge in personality based coverage or leverage the audience's in-built understanding of culture and politics, so you get broad statements about how urban professionals voted for party X but the rural poor voted for party Y.

The US is huge. The problem is efficiency of transmission. But maybe we just install 2x the solar to overcome that? When energy is so low cost we can afford to be a bit less efficient.

Both can be true.

The historical left/marxist view is essentially: look at someone's social context and where their money comes from, and you'll know how they probably think. The American ideology is: everyone is an individual and a tramp can be a philosopher.

The second is much more appealing to us as individuals, and to our sense of self-determination and independence. And people do have agency---some people succumb to despair, others don't; some work hard, some don't; some rebound from setbacks, some don't.

But if you look at large groups of people, the first one starts to have strong explanatory power. Look at US coverage of foreign elections, and that second view is almost all you see.

Yes, total failure on many counts. But when called to answer for it, you can respond like a flawed human or you can have a pr crisis firm present you as a bland corporate robot who says things like “if anyone was offended I apologize” or “we’re going to investigate how this happened.”

I think his answer was solid and seemed genuine. Doesn’t mean I think he’s got good judgment generally, but at least this was not a shitty thing to say.

"Giant Polycule" would be a great band name.

Or, you know, you could be a stand up person and be genuinely contrite.

He's a big time CEO and his company is impacted by how well he handles it. So of course there are PR/crisis people around. They specialize in making psychopaths look human.

But leaders are where they are because they make good decisions - their character does matter. Calling in the image-fixer types to clean up your mistakes is easy mode.

Yeah but to have research be actually useful, you kinda have to have a full-on frontier lab.

Or be content to publish academic papers, which is what this team was doing.

The headline makes good bait for AI skeptics. But this is a team of 50 people with a grandiose name.

The more interesting idea is that MS has abandoned AI research entirely, presumably because their zillions invested in other companies are a better bet in that area.

If you know, you know. Lots of bystander comments in this discussion, finding fault or claiming they’d do it better. Maybe NBC News should be recruiting from here.

And all that is missing the real story, which is how one guy made tech magic happen, and how it feels to be praised for a technical accomplishment when those praising have very little idea what it really took.

> devs take note

There’s been an incredible expansion of computer hardware and networking in every dimension since the heyday of the floppy. And the floppy itself was a huge step up from technology of even a few years prior.

Your comment could have been made in any year since 1950, and it would have been both correct and irrelevant outside the areas most resistant to hardware upgrades. FAA is conservative, and should be, but the rest of us have supercomputers in our pockets.

Have software developers abused the hardware capabilities they have been granted? Of course they have. I’d love to see some care put into that. But beyond the basics, hand-twiddled hyperoptimizations are now an exercise for craftspeople. Or…AI?

Read enough of these “explanations” of what the market did and a portrait emerges of extreme emotional volatility combined with a fantastic imagination. The market is insanely happy because x went up, which means the Fed will do y, and so pork belly futures will do z. Oh no! Now it is sad! But for the same reason?

My hope is that this moment is a decisive but constrained tempering of market enthusiasm. But that sounds too reasonable, so…probably next week it’ll all be back to crazy town.

Well, my reference for a commonly used incorrect definition is “anything that involves using an LLM” so yours is better than that.

But the elements that I think are necessary are 1) goal-oriented processing, 2) ability to use tools, 3) ability to iterate and self-correct. Along with that typically comes 4) self-talk or other hidden context during execution and 5) support for defined personas beyond system or user prompts.

That’s off the top of my head, probably not complete/correct, and definitely meant to be more technical than a non-tech person would need.

This is clearly voter sentiment about AI in general and big tech in general.

Most voters don’t know what agents are. Many of my coworkers in tech can’t even define “agent” correctly.

So no, the world is likely not suddenly agreeing with you about “jagged performance.”

“people” = 3 people with treatment resistant epilepsy

“nearly vanish” = 90% reduction. 10% is still a lot

Interesting, but early.

I have seen more distinct candy brands/types in those shops than I ever have in the US. Plus jars of marshmallow goo called “Fluf” for some reason?

Shhh we don’t talk about it

This is buried at the end.

> My goal here is not to write the end-all-be-all benchmark for “I’m a vibe-coder, which framework should I choose?” The point of this post is to demonstrate that it’s possible in the era of AI coding to still make decisions based on evidence rather than vibes.

But it’s obvious that the article will be understood as a useful benchmark, so author probably should have addressed that point earlier and more thoroughly.

Taking the authors pov for a moment, they just showed you how to put all your criticisms to work in making a newer better benchmark :)

If you don’t start new ones, will you really finish the unfinished ones? Really?

It’s ok to admit that the beginning is the most fun.

I must live a very vanilla, sheltered life because I’ve never bought anything and wondered if a payment processor would object to it.

That’s almost always a delayed effect. Move software people around and the revenue effects aren’t instant. In a heavily R&D-oriented area it might be 2y or more.

Exactly, that's why en-us is different than en-gb, en-ca, or even en-jp.

I would assume that a company with the resources of an OpenAI or Anthropic would have already thought about this very thoroughly, based on Spectre/rowhammer/meltdown/etc. If a random HNer (no offense) can have a hardware 101-level thought, someone there would have that thought too.

But that assumption has been wrong before.

Faring well in general, sure.

I read it as: faring well in a work/industry/employment situation where expectations for velocity and quality pivot on the fact that 99% of your peers use AI, and 90% of those use it to do poor quality work quickly.

Now your boss is surprised your email took more than 10s to produce, and wonders why you are concerned about what it contains. None of your teammates are.

Would you do this work but not expect credit for your progress? Especially if your career depended on it?

Not many people create things and refuse to sign them, and believe in anonymous contributions to humanity. I’ve had coworkers present my work uncredited with the implication that it was theirs. Not many people would be fine with that.

There are a few, but not many. Suggests to me that what you call “gold stickers are actually important.

The argument against aggressive term limits that I’ve read: there are many people making a career in politics as a whole—lobbyists, influence groups, party committees, etc. They have a huge advantage over the term-limited elected politician.

I’m not a political junkie but that makes sense to me. The critters who get your vote are just part of the system that runs the place.

Hamming it up is an overlooked parental contribution to reading. Lay it on thick! Do all the voices, act like you don’t know what’s on the next page, be shocked and surprised. Kids love it.