I don’t feel this is a good example of why LLMs shouldn’t be treated like PhDs.
My first interactions with gpt5 have been pretty awful, and I’d test it but it’s not available to me anymore
Edit: I am not having a stroke, I’m bad at typing and autocorrect hates me
I wrote model context protocol servers
I’ve definitely cut down a lot of my workload with them.
I have data between multiple external systems I need to manage, so I built some tools to automate it.
Since the LLM agent can do follow up queries against the APIs independently it excels at cross referencing between systems. This work is impossible to know beforehand, so you can’t traditionally automate it.
So what used to waste several hours a week of pulling data, searching with crap search tools, pulling data again, etc. In 12 browser tabs is now just an LLM doing a couple dozen tool calls.
That leaves me to just take the output and act on it.
I’ve done tests and it is accurate and reproducible enough. Plus I already have a good sense of when such info exists, so if it is wrong I can trivially redirect it.
Yeah I can’t check at the moment, but it has been an issue for me for a while.
I was very confused why it would fetch the weather for Paris randomly with some models, and it was even more confusing because I was learning MCP by writing a weather tool (so the LLMs were getting conflicting instructions).
I wonder if they updated the tool prompt in LM Studio
In case you don’t know, if you use tools LMS injects a bunch of shitty prompt that you can’t change, and it has specific examples like fetching the weather in Paris.
I have not checked that prompt since they added a deno based js sandbox, but I could imagine them adding a 1 shot example of a Fibonacci generator.
Hows drainage?
Also they kept the much cheaper pavers around it, which looks stupid as hell.
The new Ryzen AI chips and Apples Neural Engine (or whatever is called) have great efficiency for performance and can run strong local models.
Intel also announced they’re going this route.
I know soldered memory isn’t popular, but right now the performance/energy benefits are big — you just have to buy the premium models.
I think NVIDIA will keep doing their massive GPU toaster ovens, Project Digits was supposed to be their low energy competitor and has been underwhelming.
The local model scene is getting really good, I just wish I’d sprung for more RAM and GPU when I bought my macbook m1.
Even then, I can still run 8-12gb models that are decently good, and I’m looking forward to the new Qwen3 30b to move my tool use local.
I appreciate that they mentioned it.
I’m not particularly good at writing, I can understand why someone would ask an LLM to help them clarify their ideas. The intent here is obviously to improve their content rather than as a crutch to feed us shit. Whether it’s better than just writing this personally I don’t know.
I’m not going to let LLMs enstupidify my writing.
I’ll continue using en dashes and em dashes — they’re very easy to type on macos, iOS, and Android.
Caldav, carddav, and email are probably the only easily portable data interops.
I guess photos can be re-uploaded but that’s not easy.
Do notes transfer though? I know Outlook, Gmail/GSuite, and Apple all have notes but I don’t know if they transfer.
On macOS and iOS — is just two dashes or alt+dash
I use it often, it’s not a dead giveaway of AI.
All I’ve got is a nubby yubikey nano
More bytes for Ur buck?
You guys have Babylonian deer?
You’re not tricking me with those discount dice
I’ve been warned your tricks!
Yep, and 2x64gb RAM
What do they do with the daddy computers? All of mine only have female ports.
BlackBerry toast