snoman

Model selector at the top drops down and shows

* 3.5 Flash-Lite

* 3.8 Flash

* 3.1 Pro

* 4 Argon - New

And honestly, the versions are hurting their case. I ask myself “how is the oldest model, the Pro model?”

This kind of bungled naming and versioning is the kind of thing I expect from Microsoft.

This happens. It’s the SOP, but it’s because they already have someone doing the job on a work visa and they don’t want to have to fire them and hire/train someone else.

It’s not necessarily a green card problem (ok it is, but…); it’s more of a visa problem.

Paper isn’t ink.

Really? Pulling the ladder up after making your millions is almost idiomatic for those people.

As costs for legal work drop we might expect demand for that work to increase. For example, it may be the case that legal help becomes accessible to entire segments of the population that had no access previously.

> Certainly, a plethora of terrifyingly broad data collection greatly exacerbates the problem, but it's not a new problem.

I argue that this is a false equivalence. When the scale of something changes by an order of magnitude (or multiples thereof), it’s no-longer the same problem/circumstance.

> this yolo approach would give me constant dread.

When you’re completely ignorant, there’s nothing to be afraid of.

That actually hasn’t been my experience. For all that I dislike about LLM code, it maintains comments well.

Create a permanent underclass that is unable to access intelligent machines. That’s remarkably dystopian of you.

> Enforcing culture is not something I believe is possible.

> … the culture is driven by incentives and rewards for behaviour…

Not sure what your definition of enforcing is… but it doesn’t have to be by actual force.

Meanwhile, I have the opposite problem. An errant word or phrase and they hang onto it forever, completely corrupting all future work because they can’t ignore irrelevant (or wrong) details/errors of their own design.

To some extent, the AI may have adopted some of the expressions because they were in its training corpus enough (ie. People used these expressions before). Might be attributable to Baader-Meinhoff phenomenon instead of people copying AI.

As they say: in IoT the S stands for security.

Additionally I think there is something to the idea that they are trying to undermine foreign competition as a stall. They can’t compete economically for geopolitical reasons right now but they also can’t let American firms dominate the rest of the world in that market/technology stack.

If the government is going to gate access to frontier models from here on out, even if new releases are a step function change… which they’re not… then it may be even more comparable to what’s available with a subscription.

Considering that engineer never reliably estimated anything beyond a few days remotely accurately before… but now they can…

As one that does, it’s a difficult discussion to have with the executives. My peers look like their teams are producing more than my teams are and any argument along the lines of “but their code sucks” isn’t going to hold water. The executives care but until there’s actual impact or poor quality, it won’t matter, and it’s a lagging metric. Many still don’t care about technical debt and that’s been well understood in industry for a while.

It’ll take production incidents, impacted customers, and brand damage to make the executives start to prioritize quality over quantity again.

1. And some that are equally skilled that don’t. It matters, internally, to them but it needn’t matter to anyone else.

I wonder if there are different use cases. You sound like you’re using an LLM in a similar way to me. I think about the problem and solution, describe what I need implemented, provide references in the context (“the endpoint should be structured like this one…”) and then evaluate the output.

It sounds like other folks are more throwing an LLM at the problem to see what it comes up with. More akin to how I delegate a problem to one of my human engineers/architects. I understand, conceptually, why they might be doing that but I know that I stopped trying that because it didn’t produce quality. I wonder if the newer models are better at handling that ambiguity better.

I agree with you, but my gut tells me that a lot of people don’t know what a good outcome should/could look like and are accepting whatever it delivers.

I just sync down everything from my wife/kids’ Google Drive/Dropbox/whatever nightly to my NAS. Usability of a cloud solution, but with on-prem backup.

That’s so shockingly ignorant/reductive that you shouldn’t be surprised when people start ignoring you in technical conversations.

That sentiment always comes from people who are better at fighting with communication.

My understanding is that, if you directly assist someone violate a ToS, you can be held liable.

Bad analogy but the getaway driver doesn’t need to enter the bank to be guilty in the robbery.

Yes.

That explains why it is always/only wanting to play stuff that I listened to recently, or in 2023.

> What is sure is that a lot of capacity is been built and that capacity won't disappear.

They really are subsidizing what will be an incredibly healthy used server equipment market in a year or two. Can’t wait. My homelab is going to be due for an upgrade.

It only has to be right once. Humanity won’t end until it does.

But nobody trusts AI. Whenever I leave my circle of engineering people and am along the general public, I hear nothing but contempt for it.

I honestly think that the windows phone development experience is where Microsoft majorly shit the bed. The sheer volume of breaking changes (and the severity of those breaks) meant rewriting a non-trivial amount of your app from version to version. I know multiple developers that just dropped support for windows phone as a result.