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

Google “We have no moat, and neither does OpenAI”

semianalysis.com

Post reply on HN
Print view View original

341–350 of 1001 posts

Re: Google “We have no moat, and neither does OpenAI”

#341

Post by lukebitts »

tysam_and wrote: post #167
hospitalJail wrote: post #102

Earlier quoted context omitted.

>Midjourney is more popular (from what I'm seeing) than Stable Diffusion at the moment because it's better at the moment. Midjourney is closed-source. Midjourney is easier, its not better. The low barrier to entry has it popular, but it isnt as realistic, doesnt follow the prompt as well, and has almost no customization. SD is the holy grail of AI art, if you can afford a computer or server to run SD + have the abili…

Midjourney is higher quality by a fair bit, from my personal experience and from being near a few of the top early AI artists for a good little while.

MJ edits your prompts, you can achieve the same level of quality if you use the same prompts they do (which can be found on the internet)

Re: Google “We have no moat, and neither does OpenAI”

#343

Post by yesimahuman »

wahnfrieden wrote: post #107
tikkun wrote: post #81

The part of the post that resonates for me is that working with the open source community may allow a model to improve faster. And, whichever model improves faster, will win - if it can continue that pace of improvement. The author talks about Koala but notes that ChatGPT is better. GPT-4 is then significantly better than GPT-3.5. If you've used all the models and can afford to spend money, you'd be insane to not use…

GPT4 sucks for many use cases because it's SLOW. It will co-exist with ChatGPT variants.

Yea 3.5 is more than good enough for a whole slew of tasks (especially code), and it's ridiculously fast. I rarely find the need to use 4 but certainly if there was a usecase it was significantly better at that mattered to me, I would.

Re: Google “We have no moat, and neither does OpenAI”

#344

Post by Ajedi32 »

CSMastermind wrote: post #192

I remember I was at Microsoft more than a decade ago now and at the time there was a lot of concern about search and how far Bing lagged behind Google in geospatial (maps). After some initial investment in the area I was at a presentation where one of the higher ups explained that they'd be abandoning their investment because Google Maps would inevitably fall behind crowdsourcing and OpenStreetMap. Just like Encarta…

What if instead of Microsoft abandoning their investment they'd invested directly in OpenStreetMap? Because that seems more analogous to the course of action the article is recommending.

Re: Google “We have no moat, and neither does OpenAI”

#345

Post by ByThyGrace »

sharemywin wrote: post #94
FemmeAndroid wrote: post #41

Earlier quoted context omitted.

Charity is only a moat if it’s not profitable.

This is the timeline that's scaring the shit out of them: Feb 24, 2023: Meta launches LLaMA, a relatively small, open-source AI model. March 3, 2023: LLaMA is leaked to the public, spurring rapid innovation. March 12, 2023: Artem Andreenko runs LLaMA on a Raspberry Pi, inspiring minification efforts. March 13, 2023: Stanford's Alpaca adds instruction tuning to LLaMA, enabling low-budget fine-tuning. March 18, 2023: G…

Interesting! It's like nothing has happened on the field for the last three weeks heh

Re: Google “We have no moat, and neither does OpenAI”

#346

Post by crazygringo »

gerad wrote: post #243
crazygringo wrote: post #141

Earlier quoted context omitted.

Well yes, generally in the business world all the "good stuff", the really smart analysis, is extremely confidential. Really smart people are putting these things together, but these types of analyses are a competitive advantage, so they're absolutely never going to share it publicly. This was leaked, not intentionally made public. And it all makes sense -- the people producing these types of business analyses are wo…

Beware that there’s also a ton of bias when something is analyzed internally. As Upton Sinclair once said “It is difficult to get a man to understand something, when his salary depends on his not understanding it.” In the case of this analysis - it sounds great but it’s wrong. OpenAI has a huge moat. It has captured the mind share of the world. The software it has shipped is dramatically better than anything anyone e…

Mind share is not a moat. And market share or being first is not a moat.

Moats are very specific things such as network effects, customers with sunk costs, etc. The very point of the term "moat" is to distinguish it from things like market share or mind share.

The article is correct, OpenAI has no moat currently.

Re: Google “We have no moat, and neither does OpenAI”

#347

Post by opportune »

ChaitanyaSai wrote: post #86

This is easily among the rare highest quality articles/comments I've read in the past weeks, perhaps months (on LLMs/AI since that's what I am particularly interested in). And this was for internal consumption before it was made public. Reinforces my recent impression that so much that's being made for public consumption now is shallow and it is hard to find the good stuff. And sadly, increasing so even on HN. As I w…

There is some really high quality internal discussions at tech companies, unfortunately they are suffering from leaks due to their size and media have realized it’s really easy to just take their internal content and publish it.

It really sucks because there’s definitely a chilling effect knowing any personal opinion expressed in text at a big tech company could end up in a headline like “GOOGLE SAYS ” because of a leak.

If there is some kind of really bad behavior being exposed, I think the role of the media is to help do that. But I don’t think their role should be to expose any leaked internal document they can get their hands on.

Re: Google “We have no moat, and neither does OpenAI”

#348

Post by ngngngng »

sdenton4 wrote: post #329
ngngngng wrote: post #327

Really interesting to look at this from a product perspective. I've been obsessively looking at it from an AI user perspective, but instead of thinking of it as a "moat", I just keep thinking of the line from Disney's The Incredibles, "And when everyone is super, no one will be." Every app that I might build utilizing AI is really just a window, or a wrapper into the model itself. Everything is easy to replicate. Why…

Just gotta get to the point where we can just ask the model to code the wrapper we want to use it with...

Any wrapper that needs writing speaks to a gap in the AI's current capabilities. I just don't see why or how I would put man-hours into trying to close that gap when a future model could eclipse my work at any time.

Re: Google “We have no moat, and neither does OpenAI”

#349

Post by seydor »

eulers_secret wrote: post #331
lysecret wrote: post #144

Fantastic article if you are quick to just go to the comments like I usually do, don't. Read it. One of my favorites: LoRA works by representing model updates as low-rank factorizations, which reduces the size of the update matrices by a factor of up to several thousand. This allows model fine-tuning at a fraction of the cost and time. Being able to personalize a language model in a few hours on consumer hardware is…

If you use the web interface (oobabooga), then training a LoRa is as easy as clicking the "training" tab, keeping all the defaults, and giving it a flat text file of your data. The defaults are sane enough to not begin undermining any instruction tuning too much. Takes 3-5 hours on a 3080 for 7B, 4bit model (and ~1KWh). So far I've trained 3: 2 on the entire text of ASOIAF (converted from e-books) and 1 on the Harry…

how much memory does the 7B training need?

Re: Google “We have no moat, and neither does OpenAI”

#350

Post by wahnfrieden »

jquery wrote: post #146
wahnfrieden wrote: post #107

Earlier quoted context omitted.

GPT4 sucks for many use cases because it's SLOW. It will co-exist with ChatGPT variants.

It's about using the right tool for the right job. GPT-4 is an incredibly versatile generalist tool and a fantastic jack of all trades. However, this comes with some drawbacks. While saying it 'sucks' might be an exaggeration, I generally concur with the point you're making.

It's no exaggeration that it sucks for certain use cases where you would expect and can achieve near-realtime response, and that's fine because it's not built for that use case. I'm responding to someone saying it's always best if you can afford it
Post reply on HN