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

Claude struggles to cope with ChatGPT exodus

forbes.com

Post reply on HN
Print view View original

91–100 of 161 posts

Re: Claude struggles to cope with ChatGPT exodus

#91

Post by linkregister »

ajross wrote: post #79
nradov wrote: post #74

Earlier quoted context omitted.

That's why the frontier LLM companies are now spending a lot more to license exclusive proprietary training data from private sources in order to gain a quality edge in certain business domains.

That sounds like spin to me. If there were a clear "quality edge" in "certain business domains" stemming from "exclusive proprietary data", someone would have been exploiting it already using meat computers. But no, businesses are dumb. They always have been. Existing businesses get disrupted by new ideas and new technology all the time . This very site is a temple to disruption! Proprietary advantage is, 99.999% of…

Frontier labs are paying the same constellation of firms offering proprietary data and access to experts in their fields to train LLMs.

They are neck-and-neck only because they are participating in the arms race. The only other way to keep up is mass-distillation, which could prove to be fragile (so far it seems to be sustainable).

Re: Claude struggles to cope with ChatGPT exodus

#92

Post by furyofantares »

pgt wrote: post #77

No one left ChatGPT over that deal: they decided to try Anthropic's Claude because the Department of War gave them free marketing.

I was paying both $200+/mo and I went down to only paying Anthropic $200/mo.

My experience has, for a few months, been that OpenAI's models are consistently quite noticeably better for me, and so my Codex CLI usage had been probably 5x as much as my Claude Code usage. So it's a major bummer to have cancelled, but I don't have it in me to keep giving them money.

I'd love to get off Anthropic too, despite the admirable stance they took, the whole deal made me extra uncomfortable that they were ever a defense contractor (war contractor?) to begin with.

Re: Claude struggles to cope with ChatGPT exodus

#94

Post by politelemon »

dghlsakjg wrote: post #90
ghywertelling wrote: post #83

Earlier quoted context omitted.

Let me explain a possible moat with an example. I have curated my youtube recommendations over the years. It knows my likes and dislikes very well. It knows about me a lot. The same moat exists in interactions with Claude. Claude remembers so many of preferences. It knows that I work in Python and Pandas and starts writing code for that combination. It knows about what type of person I am and what kind of toys I want…

“Hey Claude, write out a markdown file of all of my preferences so any AI agent can pick up where you left off”

In fact, here, I'll do it myself.

Re: Claude struggles to cope with ChatGPT exodus

#95

Post by ajross »

linkregister wrote: post #91
ajross wrote: post #79

Earlier quoted context omitted.

That sounds like spin to me. If there were a clear "quality edge" in "certain business domains" stemming from "exclusive proprietary data", someone would have been exploiting it already using meat computers. But no, businesses are dumb. They always have been. Existing businesses get disrupted by new ideas and new technology all the time . This very site is a temple to disruption! Proprietary advantage is, 99.999% of…

Frontier labs are paying the same constellation of firms offering proprietary data and access to experts in their fields to train LLMs. They are neck-and-neck only because they are participating in the arms race. The only other way to keep up is mass-distillation, which could prove to be fragile (so far it seems to be sustainable).

Meh. I think there's basically no benefit shown so far to careful curation. That's where we've been in machine learning for three decades, after all. Also recognize that the Great Leap Forward of LLMs was when they got big enough to abandon that strategy and just slurp in the Library of All The Junk.

I think one needs to at least recognize the possibility that... there just isn't any more data for training. We've done it all. The models we have today have already distilled all of the output of human cleverness throughout history. If there's more data to be had, we need to make it the hard way.

Re: Claude struggles to cope with ChatGPT exodus

#96

Post by fragmede »

causal wrote: post #47
112233 wrote: post #45

Earlier quoted context omitted.

Interesting to hear! I've had completely opposite experience, with Claude having 5 minutes of peerless lucidity, followed by panicking, existential crisis, attempts to sabotage it's own tests and code, psyops targeted at making user doubt their computer, OS, memory... Plus it prompts every 15 seconds, with alternative being YOLO. Meanwhile codex is ... boring. It keeps chugging on, asking for "please proceed" once in…

> psyops targeted at making user doubt their computer IDEK what that means, specific examples?

There's this bug in Claude Desktop where a response will disappear on you. When you're busy doing many things at once, you'll go back to the chat, and you'll be all "wait, didn't I already do this?" It's maddening and makes you question your own sanity.

Re: Claude struggles to cope with ChatGPT exodus

#98

Post by clcaev »

mjburgess wrote: post #54
clcaev wrote: post #49

Earlier quoted context omitted.

Executives are certainly capable of understanding moral/ethical concerns. Around 2005, a Yale Psychology PhD candidate asked me to write a web-based survey instrument with various questions, some on complex but straight forward business questions (the controls) and others with moral/ethical aspects. Senior executives participated and they answered similarly to rank & file, often completing the entire survey much fast…

You would also need to control for the degree to which people had a stake in the outcome (ie., virtue signalling). Since executives have to make decisions where choosing the moral option may impose an economic (or operational) cost, this requires thinking through the actual choice. Morality for the "rank and file" is just a signalling issue: there's nothing to think through, the answer they are "supposed to choose" i…

I hope the addendum helps clarify.

This study showed executives spent relatively more time on questions with moral/ethical concerns. Perhaps the control questions were more similar daily work and hence familiar, while there were fewer encounters with questions having moral/ethical concerns. Perhaps executives decided more care was required for these questions to ensure people were not hurt.

Getting back to the grandparent post, executives are certainly aware of situations with moral/ethical concerns and need not consult their barber to answer them.

Re: Claude struggles to cope with ChatGPT exodus

#99

Post by bryanrasmussen »

nradov wrote: post #74
groestl wrote: post #69

Earlier quoted context omitted.

> are so evenly matched It's because the real value of the models is in what we (humanity) fed them, and all of them have eaten the same thing for free.

That's why the frontier LLM companies are now spending a lot more to license exclusive proprietary training data from private sources in order to gain a quality edge in certain business domains.

so can it be the one that gets ahead on having people go find things for them - https://news.ycombinator.com/item?id=47285283

Re: Claude struggles to cope with ChatGPT exodus

#100

Post by adammarples »

remus wrote: post #82
cmiles8 wrote: post #44

All this demonstrates how non-sticky all this tech really is. When your product is basically just an API call it’s trivial to just swap you out for someone else. As such it’s unclear what the prize at the end of the present race to the bottom is. We swapped OpenAI out for Claude and it required updating about 15 lines of code. All these guys are just commodity to us. If next week there’s a better supplier of commodit…

> As such it’s unclear what the prize at the end of the present race to the bottom is. It's a market worth many billions so the prize is a slice of that market. Perhaps it is just a commodity, but you can build a big company if you can take a big slice of that commodity e.g. by building a good product (claude code) on top of your commodity model.

The entire point of a race to the bottom is that your competitors keep reducing their prices until those billions disappear
Post reply on HN