The range of human (mis-)behavior is extremely wide, so I wouldn’t doubt that some doctors and patients are doing what you fear here. I don’t think we should form opinions on such a broad situation on the basis of a few extreme people and situations.

The question I would ask is would you rather have more people suffer from not having care, than some people suffer from receiving care that they later regret? The latter is something that’s incredibly sad, no doubt, but it’s an intractable and tragic side effect of offering major medical treatments and interventions in general; the “false positive” aspect is not unique to gender affirming care, either in its existence or its magnitude. (The politicizing of the false positive is, though, because gender in general is incredibly politicized).

Gender dysphoria solved by one-way-door gender affirming care is quite rare (there are many intermediary steps people can try and ultimately be helped with), but education about the issue and the availability of treatment helps people like my friend. I think it’s pretty unambiguously positive to universalize the availability of that care in the same way as any other form of healthcare and education, because it’s genuinely the only way some people can feel comfortable in their skin. And although there may be problems with the standard of care, the standard for care can only improve with time and experience.

For me it was a really confusing issue until I became close friends with someone whose childhood best friend is trans.

If he was born a decade earlier, he probably would have killed himself (this was the path he was on, which is incredibly tragic and all too common); the gender dysphoria invoked depression was unbearable.

Instead, he was able to work through therapy and medical care to understand his gender dysphoria and receive gender affirming care in his late teens.

Now (over a decade post treatment) he’s among the most cheerful people I’ve ever met. He inspires joy as a band teacher, is inspiringly happily married, and is raising a beautiful baby girl.

I often think about him when people talk about the issue in the abstract. The hundreds of children whose lives he’s impacted for the better, let alone the lives of his friends and family. Removing gender affirming care is implicitly saying you don’t want any of that to happen, because the logical conclusion of removing is people like him in a pit of depression and despair that often ends in suicide, all over an affliction that they did not choose.

This is where the “medically necessary” part of gender affirming care comes from.

I didn’t understand it before I knew him and his story so I don’t begrudge people who are in shows I used to walk in. But I’d encourage people to try to understand and lead with empathy and meet people where they are.

FWIW, Origin is owned by the Graphite folks inside of Cursor: the project was announced by Tomas, Graphite co-founder, at Cursor's Compile conference today.

https://graphite.com/ if you're not familiar, is a tool built on top of GitHub!

I think it’s impossible to interpret the actions of their executives here without considering this information.

Edit: It also might be interesting to point out how your solution differs from what the folks at Planetscale are building https://planetscale.com/neki

I personally don't believe in any sort of cabal (Occam's Razor hasn't let me down yet). Ultimately, I don't really care *why* they're wrong as much as I care *that* they have diverged from my rubber-meets-the-road measures of value.

That is concerning to me, because people are investing 100s of B's of capital based on the putative RoI putatively available to people like ourselves. When the benchmarks support this RoI thesis, but none of the anecdata does... that's really concerning!

Re: academics, I don't think any of the data academics have access to are good proxies for the work real people are doing. And for the data that are good proxies, the model labs certainly have access to the same data, and therefore the benchmark performance against those data is irrelevant.

This is what myself and my coworkers (and many other people in this thread) are doing on a daily basis with real stakes and real tasks – which these benchmarks are all aiming to be a proxy for. There's a real, tangible [cost]benefit to [not] using the highest-ROI models and harnesses.

The people with real incentives and skin in the game are telling you that the data diverges from "the data".

I don't mind if you don't take it seriously, our jobs are more important to us than a benchmark is.

But I wouldn't opt-out of using your own eyes and the eyes of others so easily, especially when there are literally hundreds of billions of dollars in invested capital with an interest in a certain outcome... this is how you end up in "Emperor's New Clothes" situations.

Comments URL: https://news.ycombinator.com/item?id=48390904

Points: 16

# Comments: 0

> By offering frontier inference closer to cost *and* open-sourcing everything that's sub-frontier

It's two prongs! One prong is that their frontier inference pricing is significantly cheaper/closer-to-at-cost as Anthropic's.

The subject of this thread is the other prong: offering compelling models that are sub-frontier and self-hostable.

Self-hosting models and at-cost frontier models are the high-end and low-end disruptions, respectively, to Ant/OAI/etc.'s business models.

If that inference becomes popular and valuable enough that those companies make billions of dollars in profit, those companies could use that profit to fund the building of alternative products and platforms that dis-intermediate google's relationship with the customer.

Google already has an 80% gross margin business, the biggest one in the world. Everybody wants a slice of it.

By offering frontier inference closer to cost and open-sourcing everything that's sub-frontier, they're commoditizing frontier labs' models, which inhibits their ability to durably make high gross margins on inference.

It's a strategic play.

If Ant, OAI, etc. aren't able to make 20-30% improvements on Opus 4.6 in 2026, does the music stop playing altogether? It seems like they'd lose their ability to charge >10% gross margin on inference in a span of 3-6 months.

Both parties get something they want this transaction. Alphabet gets the Berkshire halo effect and a guaranteed buyer of $10 billion worth of equities, Berkshire gets a large tranche of equity at a price they believe is fair.

I think they view Alphabet as their next Apple, and a relatively safe place to ride out whatever happens with AI: Alphabet is fairly well positioned for the upturn or the downturn, especially now with this expanded warchest of cash.

I hear and have a lot respect for what you’re saying, but I’d like to propose that we thoroughly explore every other alternative first, just to make sure we aren’t missing out on something bigger and better and leaving anything on the table.

Sigh.

But the narrative that 4.Y is an improvement over 4.X is essential to keep the model training music playing.

If 90+% of the gains come from the harness, how can you continue to justify spending billions of dollars on training and an 80% gross margin on inference on the latest model? (Reportedly what Anthropic commands on the top tier of their frontier model API billing).

So differentiating between the two (what I’m trying to do here) is really consequential!

But you’re right, the tax would have to be much more punitive to crossover into the red.

If it does make it more challenging to justify the business of being a landlord, I’m all for it though. Steps towards the end goal of more New Yorkers who want to owning their primary residence.

Edit: Actually, as a property tax of nonprimary residences, is this not also effectively also a Landlording tax? Will my landlord's tax bill go up because he's not residing in my building, if my building is above the threshold assessed value of $1mm? Or are >$1mm "multi-family homes" (significant % of housing of New Yorkers in BK/Queens) exempt and this only applies to condos?

My 2¢, I personally feel like all of the productivity gains since 4.5's release (in November 2025!!) have come from improvements to the harnesses (cc, cursor cli, codex, opencode, whatever) AND from the context window expansion from 200k to 1M.

But the actual "raw" intelligence of the model / ability to make good decisions feels like it has plateaued since 4.5. 4.6 was maybe a small improvement, but hard to differentiate from in-context-learning with the 1M window. 4.7 if anything felt like a regression in wisdom for me and my coworkers, with it consistently making worse/lazier decisions.

IMO as a github-watcher, I think they changed their definition of what constitutes a sev-0 between sev-1 for the better. In particular, they had a few "sev-1"'s around the turn of the year that would be classified as sev-0's if they happened today.

Pre-4/22 GitHub sev-1 was a normal SaaS company's sev-0, imo. So I think their new system is more reflective of reality. My guess is that a few of their big customers bullied them to have more accurate SEV categorization.