The unpredictable thing about bubbles is that you never know when it is gonna pop or the irrationality of the masses, until then it is a hot potato game.

ABS

Always Be Selling

This is true - the most compelling evidence we are in a bubble is not the content of this story (maybe it's just a day in the markets) but the triviality of the cause for hand-wringing. A somewhat disappointing product release from a single company should not strike investor dread across the entire sector. The tenor of the conversation changed dramatically over the weekend because bubbles are very thin and pop quickly.

That said, "GPT-5 will not be any better than competitors' products, demonstrating OpenAI was bluffing about AGI and destroying investor exuberance" was a very specific prediction made by (for example) Gary Marcus.

which makes it really annoying in the sense that I cannot simply short the company, because I don't know when their pants will drop :/

Easy come, easy go!

Benefits: Software companies will get back to that feature request board and stop just duck taping AI to everything with little tangible benefit to end users.

It's a tape for ducts rather than ducks :)

This malapropism has taken such a deep hold that stores actually do sell a tape called “duck tape”. It’s the same thing as duct tape, but it is labeled with the duck.

So as annoying as it is, usage continues to evolve.

Duck tape is the older usage. It was strips of canvas duck (the technical name for traditional sailcloth fabric) for finishing sewn edges and rope splices. Adhesive variants came later but fabric tapes are ancient and still in use for example twill tape. Also if you know what duck weave looks like duct tape is still obviously based on it.

It's made of duck canvas. https://en.wikipedia.org/wiki/Cotton_duck

Thank you for this. I was pretty sure most people in this thread had the duct/duck origin incorrect, and this reminded me of why I thought that. Looking at the Duct Tape wiki, it says:

> During World War II, Revolite (then a division of Johnson & Johnson) developed an adhesive tape made from a rubber-based adhesive applied to a durable duck cloth backing. This tape resisted water and was used to seal some ammunition cases during that period.

This makes me think Duck is the original term for the tape.

Made even more confusing by the introduction of "Duck brand Duct Tape".

Duck type, duck tape, duck tales!

Duck brand duct tape!

Is guerilla tape a thing yet?

I'm sure somebody somewhere at some point has to use tape on ducks (for veterinary surgery, say). That is, clearly, duck tape.

Fun fact: The product sold as "duct tape", or sometimes actually as "duck tape"(!) is not generally suitable for use in ducts.

If it looks like a duck, swims like a duck, and quacks like a duck, then it probably is a duct

Duck Brand Duct Tape. Was I really incorrect ¯\_(ツ)_/¯ https://www.duckbrand.com/products/duck-tape/utility/silver-...

Weird how people are saying bubble now and not when there was a far larger selloff between feb-april this year.

People were definitely saying bubble earlier this year, but I think the market could excuse the selloff by attributing it to DeepSeek and "Liberation Day"; the AI hype machine was still in overdrive. For some reason the market has decided not to care about tariffs anymore, but GPT-5 being mid has really put the kabosh on the whole AGI schtick.

> but GPT-5 being mid has really put the kabosh on the whole AGI schtick.

I don't think any significant conclusion, for the future of AI/possibility of AGI, should be made by looking at a single model/company.

I'm not even making predictions on whether AGI is possible or not, but the valuations of these companies hinges on AI being able to replace a large portion of human labor. I think we are teetering on the edge of a huge bubble pop unless some other company releases something that's not just another chatbot that does slightly better on some benchmarks that nobody outside of the industry has ever heard of.

What I'm trying to say is that the stock market mostly runs on vibes at this point. And when you have the CEO of the leading model company hyping up a release people have been anticipating for 2 years by tweeting out images of the death star, saying that it made him feel useless, etc. etc., and then it turns out to be a cost-cutting exercise and causes a massive uproar amongst its most loyal users, I think the market is going to adjust to reflect the bad vibes.

> GPT-5 being mid has really put the kabosh on the whole AGI schtick.

I think it’s interesting to watch this play out. I believe many people knew that gpt5 was chronically delayed as long as it could be, because the measuring stick of 2->3->4 was going to show a significant smaller step, which would inevitably destroy their market value.

It would be interesting if that does end up becoming the thing that pops the bubble. There is not longer any reason for the average person to take seriously the hype from OpenAI. They can now see for themselves that the long awaited release of GPT5 is an emperor without clothes.

Among institutional equity investors there has been very little discussion of GPT-5 being a disappointment or about diminishing returns in LLM scaling. Over the last few weeks pod shops have been pushing the narrative that vibe-coding AI startups are going to crush established enterprise SaaS companies.

I said bubble since gpt4 was hyped like the invention of bread, but I guess I'm not important enough

Bread was ok. We need the slicer!

isn't there usually a dip feb-april? I thought it was people paying taxes.

... Was anything _else_ happening during that sell-off? Was just business as usual, right?

That was a general market shakeup, this seems potentially much more focused.

related: https://news.ycombinator.com/item?id=44963715

Is this driven by AI bubble fear, or of an awareness that compute needs are finally being met, after an insanely rapid ramp?

Not much of a sell off - NVDA is down about 4% from the peak, back to where it was at the start of August and up ~35% on a year ago.

Typical stock market reporting bullshit. Nothing to see here.

On any particular days stocks will go up and down 1% - 3% just on trading and more when larger investors decides to profit take.

Drops over a certain percentage can trigger automated sell-offs creating bigger drops - and that's it. It's big players who can move the market enough to create a signal, and herd mentality from there.

Palantir for example hit a level and a lot of investors sold expecting a drop. They will now be buying back in once it hit its floor. Rinse, repeat.

Maybe the AI stocks are more volatile with wilder swings due to the hype, but there's no bubble unless they start failing or the money can move to somewhere with better returns.

Nvidia and Palantir are wildly profitable right now, so talk of a bubble is premature.

MIT study: https://news.ycombinator.com/item?id=44940944