abujazar

Separate types for each model id is an extremely tedious way of avoiding bugs that can easily be prevented by a single test.

I found iOS 7 to be a long overdue cleanup in the skeuomorphic design that had gone too far. I agree that revisions were necessary – but at least it was a step in the right direction. Liquid Glass seems like a step towards bloat and gloss with obvious usability downsides and no obvious advantages to the current design system.

Maybe, but I don't think visionOS can fully explain the move towards more transparency. Older macOS versions (in the Mac OS X days) used a lot more transparency, and the first iterations of Aqua around 2001-2005 actually had a lot in common with Liquid Glass. Transparency everywhere was abandoned for obvious usability/accessibility reasons as well as changing standards for aesthetics.

iOS 6 -> iOS 7 was a huge improvement, moving towards simplicity and minimalism. Liquid Glass feels more like bloat and a lot of wasted screen estate.

Those huge rounded corners do waste a lot of screen estate.

Liquid Glass seems like a big step backwards. It's new for newness' sake and hardly brings any UI improvements.

Agreed. I'd prefer a modern iPhone the size of an iPhone 4, it was perfegt. I made the same "upgrade" from 12 mini to 16 Pro, and the 16 Pro is so large and heavy. Feels like we're moving backwards in time.

This "attack" can't be mitigated with prompting or guardrails though – the security needs to be implemented on the user level. The MCP server's db user should only have access to the tables and rows it's supposed to. LLMs simply can't be trusted to adhere to access policies, and any attempts to do that probably just limits the MCP server's capabilities without providing any real security.

Well, this is the very nature of MCP servers. Useful for development, but it should be quite obvious that you shouldn't grant a production MCP server full access to your database. It's basically the same as exposing the db server to the internet without auth. And of course there's no security in prompting the LLM not to do bad stuff. The only way to do this right in production is having a separate user and database connection for the MCP server that only has access to the things it should.

DeepSeek actually does this is one go

By invoking, do you mean installing/configuring the MCP server? It's the LLM that decides which MCPs to use.

Will everything blow up when they create C4?

Lost me at "make an RSA key". RSA is ancient.

You can set up your IDE to use local LLMs through e.g. Ollama if your computer is powerful enough to run a decent model.

Actually you shouldn't be running LLMs in Docker on Mac because it doesn't have GPU support. So the larger models will be extremely slow if they'll even produce a single token.

As someone who's been working with cinema and video mastering, it sounds like you haven't seen the difference between professional formats like DCP and consumer formats viewed on a proper screen or projector. There's a reason we still have cinemas after all.

Even consumer equipment benefits greatly from visually lossless encoded media.

I guess the Danes (just like us Norwegians) think the choice of F-35 rather than neighbouring allies' planes was wise... In contrast to the F-16, one can only assume that Trump & co can basically disable the F-35 or at least render it completely useless for battle. Norwegian operation of the F-35 is even completely dependent on American personnel for years to come.

I believe it'd be more precise if it used an appropriately chosen and applied LLM in combination with web research – in contrast to juggling together some LLM generated code.

Nice concept – but unfortunstely I found it to be incorrect in all of the examples I tried with my Mac.

It'd also need to be much more precise in hardware specs and cover a lot more models and their variants to be actually useful.

Grading the compatibilty is also an absolute requirement – it's rarely an absolute yes or no, but often a question of available GPU memory. There's a lot of other factors too which don't seem to be considered.

Azure ≈ Klout

As if you don't have to do that when using a govt credit card for expenses.

Does this have anything to do with Trump & Musk recently banning everything non-binary and terms like "diversity" from government? I guess it was about time to fix this.

4k@32" is pretty pixellated. I'm kinda amused that people find this to be an acceptable resolution for working with text in 2024.

Natural keys can change, and synthetic keys never have to. That alone is reason enough to use synthetic keys.

Performance is another major argument for synthetic keys, as they are can be made sequential, which is rarely the case for natural keys.

Those fines are pocket money for the offenders. Why are there no real consequences?

I dont't care much for Windows, but I've been virtualizing pre-release arm64 Windows 11 on macOS since 2021.

web3 is a term coined by some of the people behind the crypto nonsense, as if it could be a successor to 'web 2.0'

«Could»? Google has already been doing this for quite some time, at least in my region (Norway), and I’d say more than half of the suggestions Google provides as top results are false.

It does enter the equation, because idle time «working» or slacking off at work could be better spent pursuing other interests that improve quality of life and most likely increase productivity over time.

UTM is awesome. I use it daily, but mostly for such plain and boring tasks as running arm64 linux VMs on Apple Silicon : )