I'd be curious if any of your customers have tried 3.1 Flash Lite. It's cheaper than 2.5 Flash, and in my experience with the free tier, quite an upgrade in terms of quality of response. My suspicion is that Google is killing off the old models because they aren't a good value for the customer or for themselves.

Whether or not it counts as social media, there is no algorithm targeting individuals as far as I know. Social media in the sense of HN is just the internet.

> that you went from those quaint, dismissive observations to a slightly panicked, "Uh Oh" realization of what these models can do?

Never experienced any kind of panic, only excitement. I told Github Copilot to add documentation to a function and it documented how the code was used even though there was nothing in the function to indicate how it was used. It somehow knew from the code pattern why I was writing that function.

> 70% complete, 10% indirect and 20% wrong

I haven't had too much problem with information in summaries being wrong, but there have been times the LLM will miss the most important details. Then when you tell it, the response is "Nice catch!" or something like that.

The key thing that makes a summary valuable is it retains the most important details while being shorter. Missing those details makes the summary wrong.

normal human reading speed is 35 WPM. Here's what that looks like, assuming 1 word equals 1 token: https://mikeveerman.github.io/tokenspeed/?rate=35.7&mode=cod...

When you say "you haven't had much problem" one can only assume you're _not actually reading the output_. In fact, like most things in modern times, one has to assume you arn't actually reading the output. You're skimming it; you're finding what makes sense and extrapolating that. This is the 70%.

The problem with non-deterministic models is that the output can't be deterministically assessed. You're harboring a delusion that you're getting real good output.

Most likely you're doing the baby extrapolation: you make it do a small, tightly scoped project and it's does 99% right. Just like a baby doubles in size in a year. Extrapolating, that baby will double again; but it doesnt.

Your human compensation limits does not extrapolate to the size and knowledge that's fed into the model and the context it extrapolates.

A strong business case for Gemma includes fine tuning, adding AI to apps that run in the cloud, strengthening Android, shifting unprofitable small AI compute to devices, and harming competitors. The first two would be done using Google's cloud services due to integration with Gemma. I think Google is currently the best positioned company to profit from AI sales to businesses over the next few years, and Gemma is a critical part of the story.

Google is actively, and directly helping companies continuously train use-case specific models based on Gemma 4 foundation. The company gets a model they fully own, trained on internal, sensitive data, and Google scoops up the profits from the training and ongoing compute spend to keep the model up-to-date.

Your comment is a slice of the reasoning underlying the "AI will take all the jobs" claim. I would constantly see references to what AI could do and how fast it was improving. Never a word about cost. We should anticipate that there will always be demand for human labor, for cheap models, for local models, and probably even frontier models.

But that's not the standard definition of "opt-in". See for instance MW: "to choose to do or be involved in something".

Yes, we know. It was sarcasm.

No idea what last.fm is. Clicked the logo at the top of the forum, but that's one of those prank links that takes you to the page you're already on. Typed in last.fm manually and got

Your request was blocked

To protect our website, our security firewall has flagged this request as potentially unsafe.

> acquire a smaller complimentary but profitable company, or start building products that compliment their current offering

I think they've been doing that, but it's tough to do it successfully. Often the best thing is to return money to the shareholders so they can look for higher returns elsewhere. I think the fact that they're still in business is kind of a miracle considering the competition.

> It is a tough market that has cut off the consumer end because all the big players have their own deeply integrated solutions: Apple (iCloud), Google (Drive), Microsoft (OneDrive).

A huge unforced error though is that the starting price for individual plans is $20/year versus $10/month for Dropbox and Box. At a certain point you have to recognize that the rules of the game have changed. Once a customer has their foot in the door with a cheaper plan that also offers better integration, why would they move to Dropbox or Box?

The typical answer when people ask why Dropbox doesn't have a cheap low tier is that the more expensive plans are more likely to be underused, and therefore more profitable than a 100GB plan that users constantly max out.

But I'm also curious about whether they've studied the long-term growth impact like you mentioned. I first needed to pay for cloud storage as a broke college student. I'd used Dropbox's free tier in high school and only needed a bit more space, and I certainly didn't have $120/year to spend on it. I ended up switching to Google Drive's $2/month plan and never looked back at Dropbox. If Dropbox had offered a comparable plan, I would have stayed and ended up upgrading to the $10/month plan when I got my first job. Looking at how much data I'm using right now, I would have become exactly the type of underutilizing user they want.

I think it makes loads of sense to be able to bring your own S3.