murkt

It triggered for me on a completely pedestrian game design prompt a couple of days ago. I’ve sent feedback and continued with Opus, but that was really unexpected

I remember similar kind of visualization from a decade ago, called paperscape. Looked cool, worked on clustering using citations and references.

Never got any idea on any use case that would be covered by such visualizations, apart from looking cool.

This supports my feelings of what works best when working with the models. Especially together with the directional bias in recall that’s mentioned here in the comments.

First, the model attention is actually limited, so less rules is usually better, but that’s common knowledge already. Or maybe it’s as common as common sense, and a lot of people still employ lots of rules and try to cram everything in one step.

Second, it’s often quite sufficient to just namedrop a technique and LLM will work differently. For example, when debugging, LLMs tend to try to brute force the problem and often end up in the weeds. Just add “use scientific method for debugging and keep journal file” is usually sufficient to improve their skill here.

Another example is refactoring. Just add “use Mikado method”, and it’s sufficient to wholly change the approach and produce much better results.

Some problems are very hard to solve with stupid code. This can easily be the case (computational geometry)

I am having a hard time picturing what could be the problem that you were solving.

Redis, Cassandra, RabbitMQ and Clickhouse. RabbitMQ looks like a black sheep in this lineup.

Yeah, probably different halves of the book. I've read it a few years ago, and it looks like I've just ignored/forgot all that "bending reality" stuff. It's not my cup of tea, so I remembered only the good parts of it

Opus 4.5-4.7 was pretty bad at it, 4.8 was a bit better, and I have not tried Fable much.

So basically you have a good enough code that’s “intuitive” for a model, screenshots, and that’s it?

I find all current LLMs to have pretty poor spatial awareness. It is becoming better, but still very poor. How are you dealing with that? Got any special tricks, any advice?

Touching shoulders and shaking hands palm-down is “persuasion” bullshit.

Can’t say anything about other books, but Never Split the Difference is about hearing and understanding other peoples’s wants and opinions. Not this pickup-artist-like bullshit.

Please read the actual book, it’s kind of the opposite from manipulation. You can read more detailed answers in the near branch of comments.

I’ve genuinely read only Never Split the Difference from your list, and it’s kind of the opposite from manipulation.

The book teaches how to actually hear people even in the very emotionally charged situations, how to properly ask them questions to understand their point of view and their needs.

If I understand my son’s needs and can give him what he wants in exchange of him giving me what I want, how is that a manipulation? I can yell at him, impose sanctions (eg no minecraft for two days) and we both will be greatly dissatisfied. Or we can both get what we want, which is a win-win.

“Never Split the Difference”, which is a book about how to successfully communicate with terrorists. Very suitable for talking with kids!

I’ve even read aloud a few chapters to my kids, because it’s very suitable for communication with parents as well

The post matches my experience as well, I am asking a question like “does A work like this and that”, and Claude responds with “you’re conflating A and B! Only A does this and that, and B does that other thing!”

Well, I am perfectly aware of B and that other thing and did not conflate them at all. I also achieved enlightment, so I don’t argue with Claude here, just ignore the obnoxiousness and move on.

Ah, okay, my bad. Got too focused on the name. Googled the dates, Satya became CEO in 2014 and Sundar became CEO in 2015, so it’s actually not that different, especially when we look at the events more than a decade later.

Sundar Pichai does not work in Microsoft, though. A bit weird to anchor the MS timeline on his position. When he became the CEO, actually? I don’t remember the year even approximately

Price hasn’t changes at all, though.

Can this be a Siri-like core? Set me a timer, tell me what’s the weather, etc. Here is transcribed text and available list of tools for the model to call, and voice the output.

The last one is about involving less people. You don't have to read it as "shut up and keep your thoughts for yourself". I read it more like "Do we really need to have six people working on this feature/present in this call?"

You can make the feature with one colleague, or you can call in five more people to weigh in and do their parts of the work.

If you involve five more, the result could be better in theory, but it will certainly take MUCH more time because of communication overhead.

What I meant is that only sometimes I am faster than Claude with debugging. When it's a standalone problem, a report in Sentry, and I just know immediately where I need to go to fix it. Then it's faster to do myself, than telling Claude what's the problem and where to look and wait.

Bugs happen during feature development, as you say, but then Claude is in the context, and I don't need to tell it where to go, it sees the bug with failing tests, or smth similar.

BTW. One thing that helps my Claude with debugging harder problems is that I tell it to apply scientific method to debugging. Generate hypotheses, gather pros/cons evidence, write to a journal file debug-<problem>.md, design minimal experiments to debunk hypotheses.

You can add that as a skill, and sometimes it will pick it up automatically, but it works wonders just as a single sentence in the input.

Sometimes debuggind is faster indeed, and making small very focused changes too.

But during feature development? Not possible. And I consider myself a very fast developer

A fee years ago, I woke up in the middle of the night with a very cool sounding riff playing in my head for a song that I was thinking about at that time. I am not a musician and that would be my first, if I would recruit enough help.

I made noises with my mouth, and it still sounded cool. Instead of recording those noises into any recording on my phone, I went back to sleep and couldn’t remember it the next morning :(

Azure is worse. These series of posts were posted here not that long ago https://isolveproblems.substack.com/p/how-microsoft-vaporize...

Autonomous Thriving Hroup?

DeepSeek pelicans are the angriest pelicans I’ve seen so far.

Frankly, that usually comes from orgs that deal with the real world. 50k generic requests per day is nothing. 50k orders per day for a small e-commerce company can be pretty overwhelming.

It’s becoming laughable when people use it to boast about microservices or something :)

When I see rps in development context, I immediately know it is requests per second. x/s on the other hand… 3x/s. kx/s looks like a physics formula to me. Some spring action? K is for coefficient, x is displacement.

writing “90 kBq” is a lot more convenient than “ninety thousand requests per second” and “90,000 requests/s”

I once made a joke during the talk that MongoDB is better than Postgres in two ways, and one of those ways is that it’s faster to say “Mongo” than “Post-gres-Qu-eL”.

Same vibe here. 90krps is not that longer than 90kBq.

With requests per minute, rpm: engine in my car revs up to 9000 requests per minute!

It’s sometimes funny to see some marketing posts like “we built our infrastructure to handle UNREAL load during the event, 100 million of requests during the day.” Which is just a bit more than 1100 rps.

In my experience, your Case 2 plays out a bit differently. "Just inherit and go" and in many cases you end up with more lines of code in total than doing the same thing in my own function with some kind of helper. Even in the dead standard thing, like displaying a list of objects with pagination. But now I don't own the flow!

Any kind of customization and I need to go jump through the hoops, I need to go look at the code what exactly happens there. But this class inherits a couple other classes, and I need to go read them as well. What for? Grug not want read seven basic multiple-inherited class just to display a list of objects.

So I disagree that it's a no-brainer. It's a no-no brainer for me.

As for writing the libraries, I have the same problems with all libraries that provide class-based API, where I need to inherit from libraries' classes to do my thing.

I like my code to be stupid and linear, to own the flow, so I can read the code and understand what happens there. Same is true for agents!

I am also willing to accept some LoC penalty for my approach. But it's shorter in practice, so win-win for me.

I was using Django since 2006 up to ~2012, and then again touched in 2014-2015. Never again.

Class-based views were a problem when they were introduced in Django, and they’re still a problem.

Especially for the so-called AI-ready framework. Because of indirection, you either have to go read all the basic classes, or read documentation three times over. Instead of just reading the self-contained view function itself, once.

Especially true for an agent, it will have to go read the new framework’s docs and source over and over and over…