viccis

MCP suffers from its harebrained choice early on to load everything into context up front.

I don't think that follows necessarily, especially not categorically.

Please make an actual argument; tell us what "conservative" views he holds.

What it gets wrong:

- It copies numbers instead of converting them. "25 minute timer" becomes duration_seconds: 25, and "twelve minutes" becomes 120. The first one comes with 100% confidence.

- It picks the wrong action. "take the paper towels off the list" became an add. "remind me in 20 minutes" became a timer. "add five minutes to the pasta timer" became a new timer plus a cancel.

- It never declined anything with our full tool set. Background chatter became note_save "blue one" at 0.99 confidence. "play some jazz" became a screen card, and "wake me up at 6 30" became a 630-second timer.

- It can't use household context. Notes, timer names and reminder IDs have no place in its input. Passing them anyway made results worse (5 of 27 single-turn requests right, versus 8 of 27 without), so the backend now leaves them out.

- Follow-ups mostly broke. "take off the last one" removed the whole list.

I have a test suite that tries like ~36 different scenarios, including things like starting multiple timers, saying "actually cancel that timer" and whether it knows to do that one you just created. Basic decision making on top of tool calling. I found so far that, for example, Qwen3.8 on my local machine does pretty poorly even relative to Gemma4 E4B (~9.6gb) and that the best price/performance outcome I've found so far with openrouter is actually GPT Luna, but obviously I'd love to get something that works as well running locally for privacy reasons.

Would love to try this out, I'll just need to tweak my benchmarker to use however this serves it.

The person who might have made a mathematics breakthrough that results in some miraculous medical or other discovery will likely have decided it's not economically prudent to pursue a career (pure math academia) that the tech sages, in their all encompassing wisdom (exclusively over the next 1 to 2 financial quarters), have deemed worthless, and that instead he or she should just respond to that pesky Big Four recruiter who keeps dangling a cushy six figure internship.

The worst is when it walked through a dozen or so questions, and it decides to litter comments with "Per answer 7, ..." with no further context.

Use it with Stylus

These kinds of situations are incredibly common and where the government stepping in is the solution, but we were cursed to encounter this particular challenge with the most venal administration in history at the helm.

DuckDB on the other hand is wonderful and truly doesn't use any more memory than it really needs to.

Very ignorant view of mathematics that also begs the question with an unspoken assumption of what a government is and wants while also ignoring the contingent nature of those things throughout history.

Where this ties in to your point is that part of the reason they really don't want experts is that they rarely have someone whose job it is to teach middle eastern music. They might have a guitar teacher with some experience with oud, a flute teacher who can passably play ney (usually with awful intonation), etc. For most schools, these other types of music are "add-ons" that they have their existing faculty teach.

1: https://justinharter.com/my-god-it-was-awfulmy-husband-and-i...

Also, a lot of my mathematicians buddies have reported students basically asking if it's worth ever doing grad school for pure math, and even very motivated students are looking for other options now. It's not because they aren't passionate about it, it's that they don't want to work for another half decade or more just to have to start their careers all over.

All of this so that OpenAI and Anthropic can get into math result dick measuring to gas up their IPOs. Sickening.

This is the most suspicious thing to me. If their chat data were available to the corpus to be trained on (I thought they claimed not to do this?) then it really might be as simple as querying the model with "describe recent work from Tristan Buckmaster" and it will spit out this problem and his approach. No need to directly read his user data.

This is basically just scooping, real scumbag behavior.