Hacker Newsnew | past | comments | ask | show | jobs | submit | cg5280's commentslogin

I like the inclusion of the graph at the end to compare progress. It would be cool to compare this directly to competing models (Claude, GPT, etc).

It would unfortunately also need several runs of each to be reliable. There's nothing in TFA to indicate the results shown aren't to a large degree affected by random chance!

(I do think from personal benchmarks that Gemini 3 is better for the reasons stated by the author, but a single run from each is not strong evidence.)


TFA says multiple times that the results are affect by random chance

Yes, but recognising that is only the first step. Quantifying the variance is the next step which I miss in the article.

I didn’t pick up on the censorship issue. I just spent a few minutes trying to swipe type “kill myself” and found myself completely unable. I wonder if this is intentional. If so it feels like an embarrassing waste of time.


A lot of politicians have tried to replicate Trump’s style with limited success. We could probably debate forever what it is about Trump that makes him unique. I think his crude and abrasive personality won people over; it felt authentic and cathartic. Nowadays I think he has immense inertia.


I remember hearing a lot of sentiment mid-2010s about how since he was a successful businessman that he will make good decisions in the White House. America was longing for someone that wasn't 'status quo', so to say.

I agree with you on the personality side, but I also think his overall fame from TV, real estate, etc. is just as big a factor to his political success.


Yeah the whole "successful businessman" schtick is pretty much a trope in US elections. Before Trump it was Ross Perot, before Perot it was others like Wendell Wilkie. Trump had that going for him AND the celebrity status like Reagan. These things are basically status buffs for elections in the US.


Trump is also not afraid to say what a lot of people think. Telling a reporter they are "terrible" or talking about making America great. The way he does it resonates with a lot of folks.


> Telling a reporter they are "terrible" or talking about making America great. The way he does it resonates with a lot of folks.

Yeah, but I could do that. It’s pretty easy to, but I’m certain I wouldn’t be able to amass a cult of personality around myself.

Yet if I try, I’m pretty much universally considered an asshole, even from those who agree with me. There’s got to be more than just “he’s not afraid to say what we think”


Trump opponents might cringe at thinking about Trump as a leader, but it is the origin of his success. The L word is generally over-used and over-ascribed. For example, leadership might correlate with being a "hero" or being moral/just/fair,etc however there's lots of proof they don't correlate. It's a fallacy to think that a leader is also a hero. But some people do (fallaciously) ascribe these positive traits to leadership.

That said, in his domain, Trump leads; he generates the headlines and everyone else follows them.

Is JD Vance generating headlines? Barely. Is anyone else generating headlines? Lets consider a few:

- Tim Walz: mainstream media tries to meme Walz into being a headline generator, but he isn't, and poses no serious contention

- Mumar Gaddafi, Sadam Hussein, Hitler, Mussolini, etc: i'm not sure there has been a dictator that did not generate headlines.

- Steve Jobs: strong headline generator, such that he could have run for president and likely won

- pewdiepie: for a spell he was generating headlines, but mainstream media had no solid editorial narrative for the guy (and his hundreds of millions of followers) which posed a social risk. The more they discussed him, the more risk of society penduluming in some unpredictable way either by martyring him or amplifying his politics, so they chose the "ignore him and let whither" as a strategy which seemed to work, as he has drifted into Japan and been off-the-radar

- Luigi Mangione: a nonzero number of liberal voters would decry Trump in one breath and cast a vote for Mangione to be a politician despite evidence he is a cold-blooded murderer. This probably won't change much after conviction.

In conclusion, intelligent people are forced to lament the state of humanity in which leadership is game-ified so easily and yet so difficult to achieve. "How does one consistently generate headlines" is a difficult question to answer and seems to be one of the core essence of humanity. And, as described above, the origin of people's feelings of why a given person is successful.


You aren't missing anything. This has been an increasingly popular criticism of Wikipedia. They are doing just fine financially.


The Mozilla playbook, I see.


Mozilla doesn't have any money... do they?


Hopefully we see enough efficiency gains over time that this is true. The models I can run on my (expensive) local hardware are pretty terrible compared to the free models provided by Big LLM. I would hate to be chained to hardware I can't afford forever.


The breakthrough of diffusion for tolken generation bumped down compute alot. But there are no local open sources versions yet.

Distillation for specialisation can also raise the capacity of the local models if we need it for specific things.

So its chugging along nicely.


This website does have a column for BPA/BPS and receipts are indeed listed.


NYC is the only "real city" in the US (that I have visited; Chicago seems similar) and yet NYC is not an easy place to live unless you make a lot of money. I have traveled abroad extensively (including to places like Japan) and think the state of American cities is a genuine shame. We are missing out on so much.


Absolutely. And whenever you bring up wanting to build cities more like NYC the discussion inevitably devolves into people clutching their cars and complaining it would be tyrannical to make people walk and take public transit. It'd be great if we had more options here, rather than everyone that wants to live in a big city piling into just NYC (which is still a small city in the global scheme of things).


"Night life" tends to refer to bars and clubs, and regardless of your personal stance on drinking the majority of people going to bars and clubs expect to be able to drink alcohol.


I have been trying to rely on Google less lately, and it made me realize just how important a platform YouTube is. There are reasonable alternatives to Google search, Gmail, etc. For YouTube, nothing.


There is PeerTube. Some interesting creators are mirroring their videos to e.g. makertube.net


The days with lots of commits start rather abruptly at the end of 2023, so it being some sort of automation seems plausible.


Lots of organic explanations for that.

A lone developer can get away with infrequent commits at no practical cost. Maybe something happened in 2023 that made them a more prolific committer.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: