porker

> Now, basically every new "AI" feature feels like a hack on top of yet another LLM.

LLM user here with no experience of ML besides fine-tuning existing models for image classification.

What are the exciting AI fields outside of LLMs? Are there pending breakthroughs that could change the field? Does it look like LLMs are a local maxima and other approaches will win through - even just for other areas?

Personally I'm looking forward to someone solving 3D model generation as I suck at CAD but would 3D print stuff if I didn't have to draw it. And better image segmentation/classification models. There's gotta be other stuff that LLMs aren't the answer to?

> They’ve dropped the ball over the past five years. Part of me thinks it was the war in Ukraine that did them in.

I'm also a subsciber for over a decade, and came here to say the same thing. I don't know how their teams were distributed across eastern Europe and Russia but the war is when I pinpoint quality declining.

I've kept my subscription for now as for PHP and Symfony nothing comes close, but I'm actively looking to move away.

> No, polars or spark is not a good answer, those are optimized for data engineering performance, not a holistic approach to data science.

Can you expand on why Polars isn't optimised for a holistic approach to data science?

> if they call their employees monkeys, certainly.

It seems to have decreased in the last 10 years but calling us code-monkeys was a common derogatory reference to the software department. I didn't like being compared to a monkey randomly bashing a typewriter but that's how things were.

It was better than what everyone called HR.

It's actually about computers unlike many of the threads today on HN?

It's an AI bot. One for @dang

Is there any way to remove every SPOF?

Currently I have multi-region loadbalanced servers. DNS and WAF (and the load balancer) on Cloudflare.

Moving DNS elsewhere is step 1 so I'm not locked out - but then I can't use Cloudflare full stop (without enterprise pricing).

Multi-provider DNS and WAF - okay I could see how that works.

But what about the global load balancer, surely that has to remain a single point of failure?

Even if you could, the DNS entries aren't loading. And then the page 404's.

Shout-out to BAML [1], which flies under the radar and imo is underrated for getting structured output out of any LLM.

JSON schema is okay so long as it's generated for you, but I'd rather write something human readable and debuggable.

1. https://github.com/BoundaryML/baml

When companies hire more senior engineers is the quality of software produced increasing and delivery time decreasing?

It's a good time to see if bodies or experience matters more.

We found worse mobile apps was good because it put boundaries around our interactions and kept us using it in a focused way during work hours.

Isn't that though the narrative arc being shaped? We see it everywhere now, but just because LLMs like to output it doesn't make the structure you're highlighting bad.

Overall I found it a decent piece, a few too many "<term>: <explanation>" blocks for my taste but better than what I can write - and than most of the tech-industry blogging I come across.

It doesn't seem to have been shared in this thread but my favourite site on this topic is https://0.30000000000000004.com/

I'm interested in knowing more about this too as I like the database localised with my app server (no network RTT!) but I also like to have a standby for HA and not being paged with everything down.

How litestream handles master-master or promoting a slave and switching back later isn't clear to me.

> when using xlsx and a good library for handling it

Which good libraries did you find? That's been my pain point when dealing with xlsx.

What did you find bad about composer?

PHP's FFI support is a feature that's 2/3 there and isn't seeing improvements. It makes me miss Python's cffi.

Yes, writing clear examples and instructions for humans was never seen as a good use of time, but as soon as a dumb mutt of an AI needs them to do stuff, they're written, and with more compassion than was ever shown to newbies asking questions...

We've lost a bunch of contributions to open source projects because these guides weren't there to help new people contribute.

Any providers in Czechia you'd recommend? It's not a market I know.

Strong agree that voice quality (and voice acting) is important. I listen to a lot of fiction audiobooks, and will listen to the end of a middling book with a good narrator, but if the narration is flat or out of keeping with the characters I'll stop after a chapter or two.

> Unless you are running some really complicated globally distributed backend your roundtrip will always be higher than 80ms for all users outside your immediate geographical area.

Many of us don't have to worry about this. My entire country is within 25ms RTT of an in-country server. I can include a dozen more countries within an 80ms RTT. Lots of businesses focus just on their country and that's profitable enough, so for them they never have to think about higher RTTs.

We have everything as scripts rather than one-liners so they work fine no matter which shell you're using.

(2018)

But can I make them wider? I don't have the precision to hit something that narrow.

(Most of the time I use the scroll gesture on the trackpad to get round this)

That tab autocomplete and predicting what I'm going to edit next is the best I've found.

The rest I can take or leave (plenty of good or better alternatives)

Is this https://insta.rs/? What are the benefits you've found with snapshot testing?

But not as bad as Outlook 365's search...

Sort-of, we've had a litany of Postgres devshops go bust over the past decade. Some will be for the usual reasons but I wouldn't want to discount the impact of RDS.

However I think Postgres is a bit different. It's a bigger market and more widely used. There's a need for tuning services no matter which platform provides the database servers, so there's a level to make money above cloud hosting.

IMO the anger people direct at source available licences would be better directed at CLAs. They're what hands the power away.

Don't contribute to projects with CLAs people without reading them carefully and understanding what can happen! Then you won't be surprised if a project is relicensed because you know you signed an agreement to let them do that.

This purist mindset is killing the long term viability of open source. If hyperscalers can make money from successful projects but open source founders can't, in 10 years we'll have fewer open source projects.

Licence freedoms don't have to be all or nothing, or apply to everyone in the world. It can be open source and have restrictions. Anything else turns open source into a religion.