FridgeSeal

Just trust us bro!!!! Just some more ai investment please! The jobs will totally come from like, the ai and stuff! We promise!! Just another billion dollars!

/s

Sure would be nice if we got to see any of that, instead of it all going solely to resource guzzling data centres selling opaque models whilst simultaneously destroying the pricing for local hardware!

LLM’s are nice. But if push came to shove, we could very easily continue on without them, or without them involved in every aspect.

Being able to write cutesy SaaS apps faster is neat, but not like, essential, the way water is.

Amusingly enough, I’ve had certain EAC games work _better_ under Linux than windows.

I used to dread seeing the EAC splash because it inevitably meant frequent crashes on windows, with no such issues on Linux.

Just get the bigger models to figure out the architecture required for hot-swappable sub-experts without loss of performance!

Got all those tokens, isn’t that the point of auto research and friends??

(Only sort of joking).

Yes I rather think “just do the Friday-Monday thing with it” will get applied, but without the realisation that once this trigger is pulled, it can’t now be “un-pulled” at a later date.

Moreover, is this a clever way for the frontier labs to try and “dodge” the negative impacts of the bubble popping.

“Oh no it’s not our fault!!1! ~~My dad works at Nintendo~~ the govt told us we weren’t allowed to release it”

> Any other scenarios?

Get to Tuesday, restriction is lifted. Get to Friday, restriction is back on. Confusion reigns.

Interesting scenario though: do the other labs attempt to “dodge” the import restrictions by claiming their models are “dumber and not a threat” thereby maintaining larger market access.

If so, doesn’t this basically force a stall in US-based development? EU will keep doing its thing at its pace. Chinese models will get a boatload more popular, but will probably slow down as they can drop to whatever pace they wish.

cynical follow-up: if they’ve plateaued, is this a clever way to avoid the negative consequences of the market implosion that a substantially substandard model release would cause, thereby giving them an “out”?

Spyware Mitosis?

If only there was a way to zoom, on your phone?!

They wouldn’t even need to do that. It’s pretty easy to come up with any number of pernicious approaches they’d use:

- “instagram is better with MetaAi: yes/ask-me-later”.

- updated ToS which bundles a “we’ll use our own ai, and do whatever we waaaaant”

Lying, gaslighting and underhanded “growth hacking” tricks are their bread-and-butter, and you can be sure that whatever they’d have you install would blindly slurp up as much as they possibly can with zero regard for user privacy.

> Do you want to give this app unfettered access to all your device data, including other apps' data?

Which Facebook and instagram will present as “tee hee updated terms of service” in the first 15 seconds, and people will tick it, because they’re not interested in reading T&C’s, just want to message their friend about dinner, and aren’t suddenly expected be deceived like that.

The physics of latency disagree fundamentally here.

Gaming, as a market, dislikes Live Service games, and many recent ones have basically flipped right out of the gate.

Stadia was the exact flop everyone predicted, and simply throwing bigger compute won’t help. That market, simply, does not want that product offering.

I get the impression nobody, outside maybe the author, has a good clue on what's going on here.

Linear.

By far and away, linear.

The task tracking features you actually use, it’s fast, you don’t waste a billion years playing “who knows how to do this and has the permissions hell” that you would with Jira. The integrations work properly. The layout and UX makes sense, and all the nontechnical people I’ve used it with liked it and had no issues with it.

Prices are already obscene, and we’re all being ripped off.

I’d much rather pay the prices corrected-for-supporting-livelihoods, than the artificially inflated prices used to line the pockets of the rich.

Fully agreed.

“Oh but businesses will leave”

Yeah so what, if they do, we either didn’t want them, or they _actually won’t_ despite the squealing, or they will go, and if their segment of the market is useful, will get snapped up by new/local versions which do respect local constraints from the get-go.

All of these are better outcomes than not doing anything because “what if”.

> The person is still willing to engage with you

But the highlighted examples demonstrate the complete opposite. Shuttling things between a model isn’t engaging with another human meaningfully.

> Lots of struggle. But not as bad as some folks are saying

I don’t know, what you’re outlining there is pretty grim. Especially as we could elect to just…not destroy people’s livelihood’s en masse?

“It’s totally fine guys, if we give billionaires even more money it will totally all work out and be fine! Everything is ok and our benevelont overlords will take earnest care of us, despite literally never having shown the merest inclination to do that ever before”.

People will just… form communities and communicate offline?

“Take the flexible, rich method of conveying information, and replace it with a slower, jankier, more limited and more rigid form!”

Sounds heinous, please never design the UX for a product I’ve got to use.

> Also, would bet money that the derived data from the meeting-summarizers is being sold to hedge-funds, to give them a bit of an edge.

And if it isn't already, you can be that they're probably to start.

All those "difficult to program but easy-if-time-consuming-for-human" tasks, will 1000% be farmed out to models at unprecedented scales.

You can’t scold them for being contrarian and wrong, and back up your point using…other comments in the thread.

The former has made a big deal about local inference and marketed that as an OS level feature.

You can also…turn it off.

Chrome silently elected people into it _and_ downloaded the model without asking because they decided that’s something they (chrome) fancied doing.

The difference should be pretty obvious.

I think you’re missing the sarcasm in their comment.

They’re saying that the emoji usage is telling them that very little effort was put into the PR and that they’ll treat it accordingly.

Sure, but that’s not an S3 concern, because the vast majority of people use S3 as it is, without needing additional management machinery.

The solution is just to spin up the machinery you need for your solution, rather than making S3 cover all possible bases.

Again, provided you either had the skills.

Or you’re getting the model to do the polishing, thereby developing no skills of your own, and we’re back to the start.

> logs, media, ML artifacts, raw dumps, etc., none of which fit into a table format.

You would be appalled at the kind of stuff I have seen teams stuff into parquet and iceberg tables.

I make the (glib) comment, because it’s a similar argument to the one that was popular a few years ago.

S3 is an object store. Treat it more like a KV store. As other comments have pointed out, the solution here is pick-your-favourite-metadata-store, be it Postgres, or what iceberg does, and other data on S3.