Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A recent experience with ChatGPT 5.5 Pro

gowers.wordpress.com

Post reply on HN
Print view View original

481–490 of 558 posts

Re: A recent experience with ChatGPT 5.5 Pro

#481

Post by coderenegade »

andai wrote: post #462
pmontra wrote: post #6

It's a very long post with a mix of technical (math) and philosophical sections. Here are the most striking points to reflect upon IMHO. > It seems to me that training beginning PhD students to do research [...] has just got harder, since one obvious way to help somebody get started is to give them a problem that looks as though it might be a relatively gentle one. If LLMs are at the point where they can solve “gentl…

> Training must start from the basics though. Of course everybody's training in math starts with summing small integers, which calculators have been doing without any mistake since a long time. Yeah, it's the same way with learning programming. LLMs can handle basic programming (and increasingly advanced programming) but I think it's necessary to write code by hand. As a beginner, of course, and arguably to maintain…

You only get good at the things you actually do. Our ancestors had to maintain a minimum level of fitness in order to be able to eat -- a level that most people today never reach, because the modern world has removed that need. Thinking is a skill just like any other, so what happens when people no longer have to exercise that skill to survive? It's a scary thought.

Re: A recent experience with ChatGPT 5.5 Pro

#482

Post by elAhmo »

Jweb_Guru wrote: post #389

This jives with what I've experienced in the brief time I had access to 5.5 Pro. It's the very first LLM that I feel like I can wrangle into solving tedious, but straightforward, problems correctly. It still makes a ton of mistakes and needs to be very rigidly guided, but it does a pretty good job of tracing its own reasoning and correcting itself in a way that the other models do not. The downside (not noted in the…

> It's the very first LLM that I feel like I can wrangle into solving tedious, but straightforward, problems correctly. It still makes a ton of mistakes and needs to be very rigidly guided, but it does a pretty good job of tracing its own reasoning and correcting itself in a way that the other models do not.

I swear that people have said the same thing with effectively every new model that came out in the last six months.

Re: A recent experience with ChatGPT 5.5 Pro

#483

Post by gapan »

dyauspitr wrote: post #49
skrebbel wrote: post #43

Earlier quoted context omitted.

Fwiw your comments here read to me as “I’m super rich and everyone I know is super rich too, and I can’t imagine that anyone isn’t”.

People spend much more than that on just commuting to work if you can spend $200 a month to supercharge what you do at work and 1000x your productivity it’s a no-brainer.

So, should they stop commuting to work to afford their AI subscription? Or perhaps just don't eat anything for 5 days a month?

Re: A recent experience with ChatGPT 5.5 Pro

#484

Post by hodgehog11 »

globular-toast wrote: post #477
hodgehog11 wrote: post #105

Earlier quoted context omitted.

May I remind you that this is Timothy Gowers. He says he doesn't understand, but he most certainly has far greater capacity than most to detect complete junk from a maybe plausible argument. His colleague is even better able to judge this, hence why he sent it to him. Also if he did send me complete junk, I would still parse it for multiple days to see what is there.

Yeah, it doesn't make a difference for me. It's the generation part. Gowers should have sent his prompts to the colleagues, not the generated paper. That's all. I feel like it's creating obligations for others to help with the remaining 20% which always takes the most time, while you get to have all the fun of doing the first 80%. I'm not criticising Gowers directly in this instance because he's exploring the possibi…

No offense intended, but this sounds like you're projecting impressions about LLM output for programming. Your words make a lot of sense to me in that context, but not as much here.

I can speak with a reasonable amount of experience here. I absolutely guarantee you that what Gowers sent over was the vast majority of the work involved for a proof. It's also an interesting exercise in general, hence the blog post.

Parsing a proof like this is _much_ easier than creating it.

Parsing code often seems like the opposite in my experience, where it is more difficult than writing it yourself.

Re: A recent experience with ChatGPT 5.5 Pro

#485

Post by aspenmartin »

foobar10000 wrote: post #444
aspenmartin wrote: post #290

Earlier quoted context omitted.

Yea there’s no logical stopping point when you use that logic. Why not say electricity or the element silicon?

I mean - I'd say electricity, agriculture, steam power, metallurgy, silicon computing (cmos), atomic power, the scientific method - these are _all_ very impressive - all lead to drastic changes for humanity. Not sure how I'd rank them. I personally think AI will end up sitting in the top 3 of these - but that is an opinion. I do think it is obvious it is at least _somewhere_ in that list.

Absolutely agree

Re: A recent experience with ChatGPT 5.5 Pro

#486

Post by fluidcruft »

elAhmo wrote: post #482
Jweb_Guru wrote: post #389

This jives with what I've experienced in the brief time I had access to 5.5 Pro. It's the very first LLM that I feel like I can wrangle into solving tedious, but straightforward, problems correctly. It still makes a ton of mistakes and needs to be very rigidly guided, but it does a pretty good job of tracing its own reasoning and correcting itself in a way that the other models do not. The downside (not noted in the…

> It's the very first LLM that I feel like I can wrangle into solving tedious, but straightforward, problems correctly. It still makes a ton of mistakes and needs to be very rigidly guided, but it does a pretty good job of tracing its own reasoning and correcting itself in a way that the other models do not. I swear that people have said the same thing with effectively every new model that came out in the last six mo…

I think it's because people walk every model up to its limits and become very aware of a task they can't make work. They do a lot of work simplifying and understanding limitations at that boundary. Then an improved model comes out and they immediately toe that barrier and make swift progress. They will also notice that the new model is natively doing tricks they had done manually.

The reality is likely that everyone is hitting similar barriers and the solutions are somewhat generalizable and get added to training new models.

Eventually people will reach the new limits and the cycle repeats.

Re: A recent experience with ChatGPT 5.5 Pro

#487

Post by robeym »

I think progressing as humans is something to be proud of. I care less about who gets credit and more about what we can now do.

I also do not think this makes people less capable of solving hard problems. The bar just moves up. More people can now work on harder problems with better tools.

If the goal is credit or proving real skill, then focus on harder problems, like ones AI can't reach

Re: A recent experience with ChatGPT 5.5 Pro

#488

Post by oblio »

fc417fc802 wrote: post #420
bdangubic wrote: post #397

Earlier quoted context omitted.

I love this optimism but I after a (too) long career I think that 3rd thing will win out - "how we do new things - but cheaper (or as cheap as possible)" there are sooooo many different articles that have been discussed here on HN that basically argue "coding has never been the bottleneck" which to me is the biggest lie SWEs are currently trying to tell themselves, I have been coding 30+ years now and coding has alwa…

> that basically argue "coding has never been the bottleneck" > we have all this work that needs to be done and not enough people to get the work done I believe the reasoning is roughly to ask, what was occupying the developer hours? Was the majority of it typing out lines of code or was it reasoning about higher level concerns? It usually comes up in response to predictions that the role of developer will be complet…

> I believe the reasoning is roughly to ask, what was occupying the developer hours? Was the majority of it typing out lines of code or was it reasoning about higher level concerns?

You're obviously right and the people who think that are the managerial types that think software developers were glorified secretaries writing after dictation.

LLM is great at generating stuff, but it's basically 3D printing. Amazing, but most of the high quality stuff in the world needs to be built at large scale out of aluminum, steel, wood, etc. Yes, I know there are large advances in 3D printing, but maybe 0.000000001% of all manufacturing in the world are done using 3D printing. A lot of stuff will probably never be possible using 3D printing.

Re: A recent experience with ChatGPT 5.5 Pro

#489

Post by n64controller »

MinimalAction wrote: post #325
hodgehog11 wrote: post #98

Earlier quoted context omitted.

As someone who is much further down the track, I would kindly suggest you drop that line of thought. I've seen far too many brilliant and ambitious people drop into depression because of it. You are worthy of doing this work because you are able to do it. Do the work because you love it and because you love the mystery. Enjoy every moment that you get to do it. Find joy in the great fortune you have to do this work w…

Thank you for this comment. I often fall into the why of graduate school many times. The pay is insufficient, hours are long, but at least I find it very satisfying on good days. It is just the feeling that what I do may not be unique anymore is what sucks. I didn't necessarily mean to find glory through incredible work alone, but through being unique in the problems I choose. Anyway, I digress.

And this matters to you? To be unique? So you care about what other people think about you and you must be special in their eyes? Cuck beta male mindset

Re: A recent experience with ChatGPT 5.5 Pro

#490

Post by DoctorOetker »

williamstein wrote: post #458
DoctorOetker wrote: post #443

Earlier quoted context omitted.

would Wiles be willing to transcribe his proof for the metamath verifier? it can be done offline indeed...

https://github.com/ImperialCollegeLondon/FLT

I asked about Wiles, because others frequently run into issues while formalizing.
Post reply on HN