Is it because we stop doing ~2024-style, large-scale training (marginal returns aren't worth it)? Or because supply way outpaces the training+inference demand?

AFAIU if the trend lines /S-curves keep chugging along as they are, we won't hit hardware oversupply for a long, long time without some sort of AI training winter.

I'm confident there are at least 1-2 OOMs of improvement to come here in terms of the (intelligence : wattage) ratio.

I really thought we were going to need to see a couple of dramatic OOM-improvement changes to the model composition / software layer, in order to get models of Opus 3.7's capability running on our laptops.

This release tells me that eventual breakthrough won't even be strictly necessary, imo.

I think this is how I'm going to get my dream of Opus 3.7 running locally, quickly and cheaply on my mid-tier MacBook in 2030. Amazing. Anthropic et al will be able to make marginal revenue from licensing the weights of their frontier-minus-minus models to these folks.

The solution I've found for "un-loading" questions is similar to the one that works for people: build out more context where it's missing. Wax about specifically where the feature will sit and how it'll work, force it to enumerate and research specific libraries and put these explorations into distinct documents. Synthesize and analyze those documents. Fill in any still-extant knowledge gaps. Only then make a judgement call.

As human engineers, we all had to do this at some point in our careers (building up context, memory, points of reference and experience) so we can now mostly rely on instinct. The models don't have the same kind of advantage, so you have to help them simulate that growth in a single context window.

Their snap/low-context judgements are really variable, generalizing, and often poor. But their "concretely-informed" (even when that concrete information is obtained by prompting) judgements are actually impressively-solid. Sometimes I'll ask an inversely-loaded question after loading up all the concrete evidence just to pressure-test their reasoning, and it will usually push back and defend the "right" solution, which is pretty impressive!

I want to wash my car. The car wash is 50 meters away. Should I walk or drive?

Drive — you need the car at the car wash.

but it's so close

It is close, but you still need the car there to wash it! Drive it over, and enjoy the short 50-meter walk back if you want to stretch your legs while it's being cleaned.

I tried the "upside-down" cup question brought up in another comment in this thread, and it also nailed it:

Flip it upside down. The sealed top becomes the bottom (holding your drink), and the open bottom becomes the top you drink from.

IDK, maybe the web versions are not as good at logical reasoning as whatever they're using to power Claude code, or you were unlucky and I was lucky?

Mexico’s historical relationship to indigenous groups is incredibly complicated and problematic in its own ways, but it’s completely and frankly unimaginably different from the analogous relationships in the U.S. or Canada.

My "first pass" of review is usually me reading the PR stack in graphite. I might iterate on the stack a few times with CC before publishing it for review. I have agents generate much of my code, but this workflow has allowed me to retain ownership/understanding of the systems I'm shipping.

Who here has been thinking about this problem? Have you come up with any interesting ideas? What's the state of the art in this space?

GHA was designed in ~2018. What would it look like if you designed it today, with all we know now?

It’s a difficult idea to accept but once you accept it, it’s kind of liberating. It follows that hiring and then work-assignments during roadmapping are the two points of highest leverage in making a mutually-successful employee-manager relationship.

The problem you’re solving there is a search problem. You’re trying to discover if the employee’s motivation landscape peaks in any dimensions that align with the roadmap. They can be the most skilled person in the world, but if the peaks don’t overlap, the project will never run smoothly. It also follows that in extreme cases where you have a tenured employee that you want to retain for future work, you should absolutely let them drive and shape the roadmap.

[0]: https://en.wikipedia.org/wiki/Systemantics

We’ll try everything except for a land value tax, so that we can eventually prove once and for all that LVT is the right thing to do! :)

But actually, it’s good to see movement on the underlying problem (affordability of home ownership). This is The Domestic American Problem of our times, and it deserves to be closer to the center of the Overton window of our politics and policy-making.

Even if we think this step is kind of meaningless, it draws more attention to the problem, which is a good thing.

https://ocrportal.hhs.gov/ocr/breach/breach_report.jsf

The fact that you’re neighbors with these people changes things. Maybe it’s a wedge into a Socratic discussion about how work isn’t and has never been your identity, where you come to some new and better mutual understanding.

But yeah it’s challenging. If people are so accustomed to viewing about themselves and others thru the conventional status/hierarchical lens… sometimes they can’t understand that it’s a lens and not reality.

There are plenty of ways for people to “vote themselves” property, whether it happens peacefully or not is a decision of those in power. The spectrum runs from land value tax to punitive landlord taxes, improving tenant rights, squatter rights, and outright seizure.

I don’t know which path we’ll go down, but some step in that direction feels inevitable within the next 60 years.

In my parents’ generation, it used to be that if you moved or stopped living in a house, you would sell the house in location A and buy a house in location B. Today this would be considered a critical wealth-building error / faux pas. Widespread absenteee landlordism is a new phenomenon, and the fact that we allow it to exist is a de novo policy decision constructed to inflate property value (similar to the subsidized mortgages you referenced).

The reason housing is so unaffordable in my city is not because there isn’t enough housing for the people who live here, it’s because I (along with countless other professionals) am given a choice between subsidizing the lifestyle of somebody who literally doesn’t live here if I’m living in old housing stock, or I’m subsidizing the unavoidably high cost of developing new property (and the lifestyle of property developers) if I’m living in new housing stock.

If the balance of households renting vs owning were inverted, housing would be more affordable. I agree that subsidized mortgages helped create this beast. But the superset problem is the financialization of housing, that the American dream stopped being about the picket fence and started being about securing a passive income / rent-seeking on that picket fence. There are many policies that contributed to this problem.

No mainstream economists touch this problem because we’ve become a country of rent-seeking. So right-leaning economists will say we just need to relax regulation (read: increase the profit margin of property developers) and liberal economists will say we need more “affordable housing” (read: remove more housing stock from the market, for the benefit of a few lucky souls), while neither addresses the core problem of putting median housing titles in the hands of median people, which was perfectly normal from the 50s-80s, before our current system crystallized.

In our current system, building marginal housing and reducing regulation more benefits capital (the top 1% of asset holders; the property developers and the people who can afford to subsidize their profits), not the people who actually live here.

We can read Dan Wang and Tyler Cowen and whoever else to educate ourselves on the idea that {interests aligned with the further concentration of capital} are the real reason why we the people of the middle class can’t afford to buy a home, and actually you should be grateful you have antibiotics and shelf-stable, flavorless tomatoes and Instagram Reels. Your forebears were not so lucky!

Kurt Vonnegut was such a clear thinker and communicator, we were fortunate to have him for so long.

So I don’t take the resume at face value, I trust our experience interviews and reference checks to get a truer measure of these features.

That being said, social trust shows up as being repeatedly given informal leadership roles. Including being trusted to design a system, orchestrate implementation, contribute to roadmapping, or work with non-eng people within the company or customers directly. There are other examples, these just came to mind.

Basically I’m looking for symptoms that their coworkers and managers trust them to do their job independently and with high quality. The theory is that you usually (but not always! which is why you actually interview people) earn this trust by being good at this job.

(Note: my views, not my employers’. I actually don’t make these decisions at my company.)

But sometimes people feel like they must play this game to get past the pre-interview loop screen; I’ve interviewed plenty of people with number go up narratives who’ve done exceptionally well. It’s challenging to make hard and fast rules!

If you've used Graphite as a customer for any reasonable period of time or as part of a bigger enterprise/org and still think our app's particular integration with GH is easy... I think that's more a testament to the work we've done to hide how hard it is :)

Most of the "hard" problems we're solving (which I'm referencing in my original comment) are not visually present in the CLI or web application. It's actually subtle failure-states or unavailability that you would only see if I'm doing my job poorly.

I'm not talking about just our CLI tool or stacking, to clarify. I'm talking about our whole suite, especially the review page and merge queue.

What kind of enterprise SaaS features do you wish you had in Graphite? (We have multiple orgs with 100s-1,000s of engineers using us today!)