I’m actually trying to move back from the Claude Code style, I feel like it’s easy to become distant from your own code, and I am feeling uncomfortable with that.
I’ve “vibe-coded” some projects and when I start to find issues or go to refactor them I don’t have that memory of why decisions were made, because many decisions were never made.
It seems like they shifted heavily to prioritizing enterprise users. Starting in the last day or two I started getting much faster responses on an enterprise plan.
For what it’s worth, I tried making a GTK4 app. I got started, created a window, created a header bar, then went to add a url/path entry widget and everything fell apart.
Entry suggestions/completions are formally deprecated with no replacement since 2022. When I did get them working on the deprecated API there was an empty completion option that would segfault if clicked. The default behaviour didn’t hide completions on window unfocus, so my completions would hover over any other open window. There was seemingly no way to disambiguate tab vs enter events… it just sucked.
So after adding one widget I abandoned the project. It felt like the early releases of SwiftUI where you could add a list view but then would run into weird issues as soon as you tried adding stuff to it.
Similarly trying to build an app for macOS practically depends on Swift by Sundell Hacking with Swift or others to make up for Apple’s lack of documentation in many areas. For years stuff like NSColor vs Color and similar API boundaries added friction, and the native macOS SwiftUI components just never felt normal while I tried making apps.
As heavy as web libraries and electron are, at least work mostly out of the box.
I've used Qt off and on, and it's generally worked as advertised. Although when drawing very short lines on a canvas way back when (~2004), it wouldn't do a great job and I had to hack in custom routines that did a much better job.
For prototyping / one-offs, I've always enjoyed working in Tcl/Itcl and Tk/Itk - object oriented Tcl with a decent set of widgets. It's not going to set the world on fire, but it's pretty portable (should mostly work on every platform with minor changes), has a way to package up standalone executables, can ship many-files-as-one with an internal filesystem, etc..
Of course, I spent ~15 years in EDA, so it's much more comfortable than for most people, but it can easily be integrated into C/C++ as well with SWIG, etc.
> I've always enjoyed working in Tcl/Itcl and Tk/Itk
In the near future I need to lash up a windows utility to generate a bunch of PDF files from a CSV (in concert with GhostScript), with specific filenames. I was trying to figure out the best approach and hadn't even considered Tcl and Tk - with Itcl you might have just given me a new rabbithole to explore! Thanks! (...I think!)
I hope it works out! It's amazing how far Tcl/Tk has come since I "had" to use it as a wrapper around an X11 window back on an SGI Irix, using Tcl scripting to interface to an OpenGL backend. I think that was like 7.3.x or something in 1994. And it was pretty cool back then already! The team around Tcl is small, but dedicated and brilliant, IMHO.
>Although when drawing very short lines on a canvas way back when (~2004), it wouldn't do a great job and I had to hack in custom routines that did a much better job.
QCanvas (or was it QGraphicsCanvas?) has long since been replace with QGraphicsScene, which is much more capable and doesn't suffer from pixelation issues.
Probably. We paid thousands / year for the developer seat in our startup, and in the end, it wasn't great. I did manage to make the Tcl/Tk event loop and the Qt event loop work together, so we could have Tk windows inside a Qt app!
I think most frustrating is the system prompt issue after the postmortem from September[1].
These bugs have all of the same symptoms: undocumented model regressions at the application layer, and engineering cost optimizations that resulted in real performance regressions.
I have some follow up questions to this update:
- Why didn't September's "Quality evaluations in more places" catch the prompt change regression, or the cache-invalidation bug?
- How is Anthropic using these satisfaction questions? My own analysis of my own Claude logs was showed strong material declines in satisfaction here, and I always answer those surveys honestly. Can you share what the data looked like and if you were using that to identify some of these issues?
- There was no refund or comped tokens in September. Will there be some sort of comp to affected users?
- How should subscribers of Claude Code trust that Anthropic side engineering changes that hit our usage limits are being suitably addressed? To be clear, I am not trying to attribute malice or guilt here, I am asking how Anthropic can try and boost trust here. When we look at something like the cache-invalidation there's an engineer inside of Anthropic who says "if we do this we save $X a week", and virtually every manager is going to take that vs a soft-change in a sentiment metric.
- Lastly, when Anthropic changes Claude Code's prompt, how much performance against the stated Claude benchmarks are we losing? I actually think this is an important question to ask, because users subscribe to the model's published benchmark performance and are sold a different product through Claude Code (as other harnesses are not allowed).
One of my favourite bosses ever was a VP who kept a bankers box at her desk and very few personal affects.
She told me she kept it there because her job was to make decisions and get fired or leave if she was wrong. She was right about so many of her choices, I would have followed her into anything. Then one day I came in and her desk was empty -- she had an apparently epic argument with the C suite and disagreed with their path so she left (never found out if that was a quit or fired). The team got a new VP, but I requested to be moved to a different team as I wasn't aligned with the new vision.
When you get to a certain level part of your job becomes owning the decisions and getting fired.
I worked in retail many years, including doing store shelf tear downs and replacement and night shift stocking.
Back in the day we would get our planograms from HQ, then we’d print out all the labels on perforated paper, and walk the shelves moving product and updating the price tags, throwing out the old. The epaper tags are very clearly an improvement to that process in both time and waste. We would also check the prices using a Motorola price gun and do our fixes manually and then print out new tags or update the counts.
I’m surprised these tags are just IR blasted with no security. I would have expected they’d need some sort of code and you would simply save the code on your gun, pop a tag in front of a product, scan the product, then pair the tag all on your price gun in like 3 actions.
I also would have thought in these days we’d use Bluetooth beacons to triangulate the shelf slot too so that HQ could have a realtime map against their planos (it was not uncommon a product’s size would change and the layout would have holes or products that don’t fit on your real shelf).
Anyways, neat project! Triggered a walk down memory lane for me.
Previously, a criminal could just print their own shelf tags. They'd probably do this somewhere other than in the store to get the details right, but it was doable. (We've all probably seen rolls of blank shelf tags sitting around at the store, and thermal printers are inexpensive. So what if it's two crimes instead of one?)
And then, in the store, they could just switch out the shelf tag(s) and try to play their little scam.
Now with this new development, a criminal still needs to get the details right. Like a blank paper tag, the little screen is also a blank slate. It's just eraseable and rewritable in-situ.
The scam is the same. It's just shaped differently.
---
I do understand why the tags are simple to write. Maintaining some kind of revolving, PKI, or multi-factor auth would be harder than doing nothing, and probably slow. Fixed, basic auth would just get leaked (probably first by Home Assistant tinkerers who find some discarded electronic shelf tags somewhere and want a new display for their house).
One-way jnfrared is cheap and low-power compared to anything with RF. And resets would be a pain in the ass if things were forever associated with a certain product, or a certain place in the store.
The way it's implemented now, on reset (yay new planogram!): All the tags get pulled and put in a pile.
And then: One by one, they're removed from that pile, put on a shelf, and programmed.
That's fast and flexible, and therefore inexpensive. Inexpensive is good. If there's one thing that all retail establishments hate most, it is their labor expense.
It does fail to prevent obvious-scam from happening. But it'd probably cost more to do it "right" than to eat the losses when the scam actually works.
> Fixed, basic auth would just get leaked (probably first by Home Assistant tinkerers who find some discarded electronic shelf tags somewhere and want a new display for their house).
You know what, that is a great idea for a project of mine, where I want to display outside temp and weather forecast in the hallway next to the wardrobe. I have been musing about it for a while now: how to make it small and not stand out, how to handle power delivery, etc.
I was already leaning towards eink, and if I can get one of these price tags cheap plus hide an IR blaster in a corner that would be ideal. All controlled by Home Assistant of course. I'm going to search the usual Chinese online marketplaces tomorrow.
Just look on eBay. It's full of used electronic shelf tags, sold in bulk -- usually, with prices still on them. :)
The sellers don't know anything about how they work so it will take some digging to find the right ones, but having to dig a bit is normal for eBay (or Aliexpress, for that matter).
I had a look and knew they seemed to be about £15 here, I couldn’t easily find second hand ones in the uk (though they’re not uncommon at shops). For £40 I can get a 7.5 inch black and white screen setup (trmnl byod xaio https://www.aliexpress.com/item/1005009532501677.html)
Lots of the tags I see though do have Bluetooth or maybe WiFi for updating as well.
I do really like eink things, I want to setup a nice 13 inch one which is now more like £160 so becoming more realistic for my to buy for fun.
I’m going to have to look more into these tags because if there’s cheap second hand ones they’d be awesome.
Yes it does, unlike before, a shits-and-giggles attacker could change all the tags in an aisle into "you're gay" without showing anything on surveillance cameras.
He wouldn't gain anything but the store would lose.
But actually: It's not that broad. It's still mostly one at a time, ish. Changing a lot of them would stand out if anyone were paying attention.
Although it could certainly be broadened...but an IR emitter that's skookum enough to reliably hit all of the shelf tags in an aisle at once would probably show up as an intensely-bright purple floodlight on the cameras. That would stand out quite a lot. :)
Well it does need to read individual tag barcodes, so it is indeed one at a time, still you could make an automated camera+beam device, hide at chest level, walk through an aisle as if looking for something, then pick up something in the end of the aisle and go to checkout.
I don't think anyone is paying attention anywhere near enough to pick that up. Additionally, one could read some barcodes and make quite cheap battery-powered narrow beam emitters to be placed in store aimed at particular tags that would only power up once a day at a random time.
If the lulz are the point, then: Just build hide the thing in the floor-cleaning robot. Include a decent camera (they're very cheap) to catch the barcodes.
If the comms last long enough as the machine passes by to program some tags every night, then some tags get programmed every night. Nobody will pay attention to the robot's new purple floodlight in the cameras.
> Previously, a criminal could just print their own shelf tags.
Between your 'previously' and now is a period of at least two or three decades, where shelf tags have only be for your information in the store, while the real price came from computerized POS-Terminals with attached barcode-readers. Which of the two has priority for the customer may depend on country, law, store policy & good will.
Furthermore stores are completely cam covered nowadays, so much luck with being seen fumbling with your gadget in front of that label, or being seen on 'tape' putting another one over it, or things like that :-)
I suppose you're right, even with auth there a risk the store would lose it and the tags would need to be trashed or manually reset. Stores aren't the most tech savvy of businesses.
On that note, I worked at a store where each department would borrow the store floppy and copy their weekly orders+sales onto it, then the store manager would read them on the office computer in one big excel file. This was in the late 2000s, so floppy was already pretty outdated, but it worked so it never changed.
I got called to the managers office one day. The store manager and assistant manager couldn't get the master spreadsheet to save and needed my computer skills. I spent like an hour fiddling with things before ejecting the disk and realizing they had hit the write lock switch on the floppy, so the spreadsheet wouldn't save.
I was the hero of the day and the manager bought a thumb drive from the photo department so this wouldn't happen again.
>Previously, a criminal could just print their own shelf tags.
Yep. For the physical "hackers" among us, a price sticker gun (those little orange or white stickers with a number on them that mom and pop shops use) was one of our first tools to mess around with
I’m sorry but this entire premise is dubious at best hilarious at worst.
I’ve worked retail on and off for a decade and been friends with AP in most places and no one has ever mentioned this happening. Never been told to watch for it, or heard a rumor about it from another store.
Just recently I was in a small shop where I was surprised to see epaper tags and ended up talking to the owner about them. She said they were super flaky and would reset at random. Agter that interaction I am not at all surprised a flipper could mess with them. But I also have not seen them widespread at the physical outlets I shop at.
reply