viccis

As a platform team member:

"almost never a product manager handing you roadmap"

Oh well wake me up from my dream vacation.

Absolutely. The definition of "harmony" doesn't begin and end with functional harmony.

Rich can easily be given with an example rather than purely aesthetics. Coltrane introduced more advanced harmonic movement via "Coltrane changes" as one example. It had been done in one-off cases in classical music (like just about anything), and occasionally in show tunes (Have You Met Miss Jones) but Coltrane was the first to isolate the specific pattern and make it into a reusable substitution technique to reharmonize other tunes.

Aside from that, Coltrane's contribution to modal jazz is significant. Though not singularly from him (or anyone else), he contributed quite a few standards and recordings to a style that was a pretty radical and rich departure from functional harmony. And if you don't consider the harmonic system this resides in (Russell's LCC is a common one used to analyze it) as "rich" then I think you might be too harsh a critic.

Also a jazz musician and I find that it depends on what time frame you're talking about. I like him a lot up to the very early 60s, and then he loses me around this point. I prefer the direction other hard bop players took at this time. Coltrane benefited greatly imo from Tyner's rock solid comping and Jones' aggressive drum work. His sax soloing got a bit too interested in dense excess, often fairly inscrutable even before he drifted into free jazz. Opinions absolutely vary on this point though. I think it still sounds good even if it's not what I reach for.

>an entire school of derivative jazz players had long since developed upon and retroactively made his innovations feel far more common

The worst of this to me is at straightahead jazz jams when sax players run out of ideas or lose the form and then just start honking lol

>And "rich" harmonies doesn't really mean anything.

I can't even begin to comprehend how you can't understand that someone can write harmonically rich music.

It's crazy that we went from "we found a significant number of students used AI in our course and gave them a 0" to "yeah they all use it we can't stop it". Basically just accepting widespread cheating, normalizing and accepting the complete abandonment of the first lesson in ethics we are taught and held to in academia.

I know this is absolutely impossible because colleges and universities have headed in completely the opposite direction, but the solution: Literally just expel them. Just like they did when people cheated when I was in college. Don't make it a small punishment so that they turn it into a risk/reward assessment. Colleges are overpopulated anyway; just trim the fat of students who shouldn't be there. Do a follow up oral exam interview about their project and if they can't explain what someone like me spent 30+ hours building and can speak about for 4 hours if asked, then expel them. They can make your coffee at the nearby Starbucks, who cares.

If you don't do this then understand that you are telling your students not using AI that they will need to begin cheating like the others to be able to keep up with the changes. OP mentioned they massively increased the scope of their assignments, which means OP is forcing their students to use AI now.

If you ever played Monopoly, you were intended to learn a bit about it whether you knew you did or not.

I don't use Plan mode because I plan out work with a ticket based approach (that also works as a memory store), but I wish there were some kind of Plan mode like harness feature to try as hard as possible to stop it from beginning to code. With the newer models, I have started to have a lot of trouble with them proceeding through the QA with me where I scope out the work and refine requirements and then immediately begin to do the work rather than either making tickets or giving me the full plan to thumbs up like I tell them to do.

So it sounds like we use it somewhat similarly, just taking a glance at it before the work starts, and that's becoming harder to do in my experience. It is important because problems with the overall plan or strategy end up magnified the further you proceed with development.

Remember that 20 commits ago is like 12 hours ago

I actually ended all govt contracts I was working on after this and ended my clearance forever. This was my "I can't keep working for them" moment. One of the purest examples of the US operating on ISIS levels of morality that exists.

Already happens in my experience. I'll try to figure out why it keeps doing this one thing and it turns out it's from some poorly advised info it put in a markdown file 20 commits ago that CLAUDE.md or AGENTS.md tell it to treat as gospel.

So it beats Fable 5.1, by quite a bit, on every metric? Interesting.

Might have to use my $20 Claude sub some more. I was moving away from it to a $100 OpenAI one to avoid the Claudese and poor token efficiency of Opus 5, given that I couldn't use Fable 5.1 with my tier, but this is worth trying out.

People often use a bunch of subagents, poor context management (though Codex's tight context limits and constant compaction tend to mitigate this), a bunch of projects at once, etc., as well as not using workflows that do heavy planning once up front and then consult it rather than thinking endlessly about what to do during the implementation part.

It's also the case that working on massive codebases is just a different beast. If they've been slopmining a monorepo for months with 200x, then their codebase is probably Lovecraftian at that point and requiring extensive effort to iterate on.

>But... data centre work is horrid, dirty, hot, and underpaid for what it is

You left out "loud" haha. Just the nonstop buzz; it's like you're working at an airport tarmac or something.

That's where you are completely wrong. The point of MCP is that you can have an agent and a harness without running raw CLI or Python commands. Very common for relatively lightweight loads that involve shuffling data around between APIs, often run in a tiny serverless task.

Honestly just getting it integrated into mobile phone swipe keyboards would be a godsend. If I type "We need to get going " and then swipe the word "now", I really do not think "mower" should be the word it chooses. Present a set of swipe-based likely words and the preceding text message to Jev, or similar model, and pick its highest prob word.

Watermarking has referred to this "spy" use case for quite some time. Digital items purchased for download often have them, for example. Even before the rise of digital downloads, screeners for movies had them.

Yeah I use my 256gb Mac Mini for iOS dev, which has always had enough space for an iPad and iPhone simulator with plenty to spare, but this eats way into that.

Not really.

The source of that talking point was a young black protester who wanted real investment into black communities to stop the biggest killer of black Americans: criminals in their own communities. He wanted a structural solution rather than a media circus about an issue (police violence) that is minuscule in comparison to structural racism and would be rotated out of the conversation as soon as the media moved on.

Greenwald's main point seems to be that the conversation about race was used cynically to attack one of his coworkers for completely spurious reasons, and I don't see anything false about that.

If anything, ignoring the cycles of crime and poverty that black Americans are disproportionately victim to in favor of occasional media spectacles is completely on brand for America's dominant right wing neoliberal ideology.

Yeah but you're still assuming it's running on someone's machine with a CLI to even use.

What if the agent doesn't have a CLI?

Yep, exactly. People who aren't super familiar with American propaganda often don't appreciate this enough, but every stupid network here will feature you if you pissed off the other team, and when Biden's necrotic ass is overseeing our waning genocidaire state, anyone who speaks out against his baby annihilation tendencies is going to end up having to slum it up on Fox because they're desperate enough to feature you.

Exactly. A lot of people complaining about MCP are doing so because their only interactions with LLMs are via big batteries including code harnesses and don't understand what kind of (usually much more domain-specific) agentic systems are being built. For example, "CLI vs MCP" doesn't make any sense whatsoever if the agent doesn't have access to a CLI!

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

>If I build a business I get to decide what to do with it.

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

I'm not the biggest fan of him, but him showing up on conservative outlets because the other ones wouldn't platform him (due to going against US Dept of State propaganda) does not mean he is "conservative".

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

It's not a "false equivalence". Multiple people are trying to explain to you what "stateless" actually means and how's it's materially different than being a fugitive, but you are refusing to listen.

Thanks for responding. I ran it just now and it looks like it could be useful if I change and limit the scope of the kinds of actions I need. Here is a breakdown of where it struggled vs some local models (~8-12B parameters running on a 16GB Mac Mini or my desktop's 3080), bear in mind I had Claude integrate it into the benchmarks and this is its interpretation, not my own:

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'll try to get something set up to try this out. I've been working on an ESP32 based Echo replacement that sends audio back to a backend server I run, and one question I had was whether models small enough to run on a Mac Mini or even smaller hardware are good enough to handle basic tool calling functionality with a bit of reasoning where needed.

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.

There's been some knot theory applied to molecule analysis chemistry as well as things like topological quantum field theory, with both of those being examples of fields that benefit greatly from previously unapplied mathematics introducing tools to use.

Depends. Do you think human mathematicians will be entirely useless for any kind of input into the act of doing mathematics? If not then we'd be advised, imo, to keep funding which is inherently just a training pipeline.

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.