IS2511

It's hard for me to fully comprehend that there are people who believe LLMs are strictly useless in programming and/or strictly harm the quality of code/product. Makes me want to write a blog post about it... Various scattered ideas and notes off the top of my head (*not fact checked*):

LLM as a smart(er) auto-complete. No, I'm not joking. This is about the level of responsibility it can handle (pretty "ok" at it). Propagating a small change with a clear pattern? Do once, TAB for the rest - an actual time save and convenience. Is that vibe-coding? I don't think so. LLM use? Yeah.

It's hard for me to fully comprehend that there are people who believe LLMs are strictly useless in programming and/or strictly harm the quality of code/product. Makes me want to write a blog post about it... Various scattered ideas and notes off the top of my head (*not fact checked*):

Of course, if the only thing that matters is shipping that code, to maximize shareholder value, and the business is pushing for "vibe-coded now, fix later/never", then it does make sense to vibe-code. Sacrifice ever achieving good for quickly achieving a working meh most of the time.

It's hard for me to fully comprehend that there are people who believe LLMs are strictly useless in programming and/or strictly harm the quality of code/product. Makes me want to write a blog post about it... Various scattered ideas and notes off the top of my head (*not fact checked*):

And model context plays into the quality too. Asking it to continue low quality text obviously makes low quality output more likely because in training data low quality text is usually surrounded with more low quality text. Having a bad codebase nerfs a model too, no matter how hard they "align" it.

It's hard for me to fully comprehend that there are people who believe LLMs are strictly useless in programming and/or strictly harm the quality of code/product. Makes me want to write a blog post about it... Various scattered ideas and notes off the top of my head (*not fact checked*):

Average quality of all code ever? Even just all code on GitHub? It's shit. Not extremely bad but pretty damn bad. An LLM can only reproduce that, so the output is obviously low quality code too. Power law, few popular repositories with high quality code vs many many low quality repositories, etc...

It's hard for me to fully comprehend that there are people who believe LLMs are strictly useless in programming and/or strictly harm the quality of code/product. Makes me want to write a blog post about it... Various scattered ideas and notes off the top of my head (*not fact checked*):

So no matter how long and hard you train an LLM, the average output quality is going to hover around the average quality of the training data, maybe slightly above. Model weights are based on frequency (read "popularity") of input text (token chains) (simplified).

It's hard for me to fully comprehend that there are people who believe LLMs are strictly useless in programming and/or strictly harm the quality of code/product. Makes me want to write a blog post about it... Various scattered ideas and notes off the top of my head (*not fact checked*):

After reading the documentation manuscript, peeking at other configurations and lots of trial and error I think I might have finally stumbled into a general working zapret2 strategy. It's my 3rd or 4th try, yet another fresh look finally got me there. The manuscript btw: github.com/bol-van/zapr...

Still, it appears to not work for Telegram for example. A *general* strategy, against *general* blocking methods. I guess for Telegram I would need something more custom... Probably not worth it since my main zapret2 instance is for LAN, but Telegram's main use-case is communication from anywhere.

I think I figured out a fix (pending confirmation). I'm not sure what is causing the link downs, probably some hardware reliability issues at 10G, but I'm fine with it as long as it can recover, which it failed to do. The fail state was link up with only an IPv6 address. I present to you the fix:

Maybe I should set a static IPv4 address... The DHCP lease is static anyway. The IPv6 recovery seemed quick (SLAAC perhaps), especially compared to an IPv4 re-acquisition, probably bc of all the DHCP yap. I think the instant IPv6 re-acquisition made it skip IPv4, causing permanent connectivity loss.

It was not in fact the motherboard port... :( Ok, the warnings are because of the WinBox itself. But what is up with all the link up and downs? This is after switching to an SFP28 port on a ConnectX-4 PCIe NIC.

It works now so I give up. Until this problem comes up again it might as well not exist.

It was not in fact the motherboard port... :( Ok, the warnings are because of the WinBox itself. But what is up with all the link up and downs? This is after switching to an SFP28 port on a ConnectX-4 PCIe NIC.

(a day later) I was drafting this late yesterday and shelved until today. Now, no issues at all for like 12h+ already. Was it software? Was it an SFP module? I don't think anything changed besides the weather outside... Anyway, now I switched to an SFP28 DAC for better reliability, just in case.

It was not in fact the motherboard port... :( Ok, the warnings are because of the WinBox itself. But what is up with all the link up and downs? This is after switching to an SFP28 port on a ConnectX-4 PCIe NIC.

I googled what could it be: 1. Spanning Tree Protocol (STP). ethtool | grep STP + cnping, compare, timing doesn't seem to match. 2. Failing auto-negotiation or bad cable. No? 3. Faulty link-state polling. Nope, dmesg is quiet. 4. Duplicate IP / ARP cache. Unlikely, skip. Filler ~~episode~~ post.

It was not in fact the motherboard port... :( Ok, the warnings are because of the WinBox itself. But what is up with all the link up and downs? This is after switching to an SFP28 port on a ConnectX-4 PCIe NIC.

> CSS/CRS309-1G-8S+ - supports up to 4 simultaneous S+RJ10 modules. help.mikrotik.com/docs/spaces/... Uhh, whoops. I wonder if this is why I get occasional connectivity losses...

Switched the cable to the Mellanox NIC, instantly feels more stable, poking the cable no longer causes instant DC. I didn't particularly trust the onboard 10G mobo NIC before, and trust it even less now. Man, enterprise/server hardware is so good... Much better than most consumer stuff.

> CSS/CRS309-1G-8S+ - supports up to 4 simultaneous S+RJ10 modules. help.mikrotik.com/docs/spaces/... Uhh, whoops. I wonder if this is why I get occasional connectivity losses...

Nvm, seems like it's just the PC having problems. Consistent minor trolling from the 10G Ethernet port on the motherboard...

GitHub internal source code stolen? I should've migrated a while ago, but oh well. Better late then never I guess. giveupgithub.org Found this cool blog post lord.io/leaving-gith.... The "i wrote it myself last week" got me, lol. At least check out "Appendix A: Table of GitHub alternatives".

Gerrit is... weird. The unit there is a "change", opening a repo shows just those, not even a current file list, usually a link to the repo on GitHub. And as @lord.io (what a profile lol) said in the blog post, the web interface is "objectively ugly". I would call it bare-bones and email-esque.

GitHub internal source code stolen? I should've migrated a while ago, but oh well. Better late then never I guess. giveupgithub.org Found this cool blog post lord.io/leaving-gith.... The "i wrote it myself last week" got me, lol. At least check out "Appendix A: Table of GitHub alternatives".

The obvious migration target is @tangled.org, but what about private repos? Well, why not remember what git remotes really are? Just SSH into your "remote" host, `git init --bare`, and it's done. You can `git remote add` it as usual on your local machine. Only works for personal projects of course.