I worked at a company that had to deal with EY as a part of a tech certification. Tha auditors barely knew why they were being sent over. Out managers and techs had to explain the process to them and assure them the numbers in the report were correct. The auditors happily accepted that, and then charged around 100k. Top job.
HN user
nimbix
The scenario I was referring to is one of many gems found in app we outsourced.
The piece of code in question had a very straightforward task: look at some bytes in the input and produce a string label to be stored alongside the whole input value. There are 5 different labels tied to equal number of fixed byte sequences.
I would like to think that most people would solve this problem using an if/else or a switch statement inside a function. Instead, what we got is a group of matcher classes, a mapping of matchers to enum values representing the labels, another mapping of enum values to actual strings, and a class that actually calls those matchers and does the mapping.
I really hope this is not the DDD way and instead we just managed to find a team that's prone to massively overcomplicate solutions to simple problems.
I recently got into c# and it seems really nice, but the "customs and traditions" of the dotNet world horrify me (eg: never write a function when you can write three classes instead)
I gave a UX person a snippet of debugging code to paste into the console, but didn't notice that the chat app ate the quotes around a string. GPT was able to fix the snippet for him. SO far this is the only case where someone I know managed to use it for something useful in practice.
I asked it to write a trivial c#/dotnet example of two actors where one sends a ping message and the other responds with pong. It couldn't get the setup stage right, called several methods that don't exist, and and had a cyclic dependency between actors that would probably take some work to resolve.
Event after several iterations of giving it error messages and writing explanations of what's not working, it didn't even get past the first issue. Sometimes it would agree that it needs to fix something, but would then print back code with exactly the same problem.
Does a device even know that it's connected to a ded cable? I'd imagine that it's only after you plug in the second device that they both suddenly see they're being connected to something.
Butt where would you hide a chess engine? /s
They made a suite of replacement tools under the name of "Adobe Edge", then they killed those as well.
Isn't XD pretty much dead? The last release that wasn't just bugfixes happened a year ago, while before that they were pushing out something new every month or two.
Domestic or international? I flew with United transatlantic around 10 times and after hearing nothing but bad things about them online for years, I was quite pleasantly surprised by how much better they were compared to companies like Lufthansa, for example. But maybe it's just me... I also loved Air France, another massively hated airline.
There was a point years ago when I was surprised to find out that you can throw things other than strings...
Yep, the "magic ingredient" is definitely insulation (along with triple pane windows).
If you pair a heat pump with good insulation, the effect can be quite staggering. My parents moved into a new house 6 years ago and their annual heating bill (for space and water) dropped from ~2500EUR down to around 300EUR.
This is in an area where night-time winter temperatures hover in the -5°C range, occasionally dropping to -10.
The old house was around 30 years old, with fairly thin insulation (5 or 7cm was pretty standard back when it was built), and had an oil burning furnace for heating.
IMHO 4k was the worst thing to ever happen to monitors. Now we have screens ranging from 24" to 43+" all using the same physical resoluton. This means that at all the most popular screen sizes you need to use some weird 1.5 or 1.25 scaling factor which results in all sorts of weird rendering quirks, like 1px gaps when rendering engines try to fit elements to pixel boundaries. Plus SVGs often rasterize to a blurry mess since there is no rendering magic that could make a 1px line rasterize nicely over one and a half pixel.
There was another somewhat similar incident in Europe today: an autonomous plane - presumably coming from Ukraine - crashed into a park in the capital of Croatia.
Many years ago I watched a documentary where it was said that pale mint green was often chosen for control panels because it was deemed it had a calming effect on the operators.
Cypress was such a pain that I simply wasn't writing or running tests just to avoid it. It was incredibly slow, had major bugs that are till open after multiple years, and worst of all, in headless mode it was passing tests that should have been failing (and were failing in headed mode). The only reason I discovered it was because headless mode seemed to run unusually fast.
I switched to Playwright 6 months ago - I'm much happier now and the switch also allowed me to delete 3/4 of the helper code that I needed before.
I've been using jEdit for probably around 18 years until around a year ago when I fianlly switched to VSCode because, unfortunately, it just couldn't keep up with all the modern development in the dev tools space.
There are two features I really miss from jEdit:
* "dumb" autocomplete that will include any word in any open buffer, even if it's a different type (html/css/js). VSCode kindo of has word based autocomplete, but expected matches are just not there on third of teh time.
* HyperSearch (search result in the sidebar) when searching in the current buffer only. So much better than having one search mode for single buffer, and a completely different one when serching through multiple files.
Isn't AutoLayout in Figma just a partial implementation of flexbox, an therefore already exist on the web in a more complete form?
I remember building those on my grandma's farm ~30 years ago. The adults woud throw hay to the top and us children would pack it down by walking on it in circles around the pole.
Another thing she would store for the winter were whole corn plant stems, arranged in a teepee-like structure around a tree trunk. The empty space between the trunk and the corn made a favorite hiding place for both us the children, as well as her chickens.
The company I worked for had a pretty much identical experience with Unilever: they aggressively reduced the scope of the contract, but expected all the extras anyway; then at the end of the contract made the company choose between getting paid for all the overages, or extending the contract. They never even really needed us if you ask me; we were primarily used as a tool for one group of execs trying to one-up another group who was backing a competitor. Eventually it all lead to us having to integrate with that competing company... until said competitor suddnly went out of business (helped by their Unilever deal, no doubt). Of course the management's objective was keeping them as a client at any cost because having them as a reference would open up so many opportunities (it didn't really).
It will come as a surprise to many people, but it's quite possible to lose a ton of money on a toxic client who is paying you millions. Avoid such "white elephant" clients at any cost. And really, the last thing you need in your life is getting dragged into some huge conglomerate's internal politics.
I never worked with Singapore Airlines, but people who have experience with them tell me they're even worse in the sense that they allegedly have a policy of not paying for software - you get to give them free stuff in exchange for being able to list them as a client. Huge software companies can afford to do that; your growth stage startup probably can't.
Some people think TPB is a comedy. Others know it's a documentary.
I definitely ran into some serialization issues when trying Piscina a while ago. Returning generated images from the workers would keep getting slower and slower until the overhead of data exchange was 20x the time it took to generate it.
I discovered through trial and error that returning image bytes directly would incur this ever increasing overhead, btu creating a binary buffer and returning that would not. It was just a pet weekend project, so I never did discover what was causing the issue.
I watched a youtuber go through the whole flow of producing a video using Descript and it doesn't do anything in terms of real video editing. It's useful for removing bad takes out of a longer video, but then you need to export the chopped up product and spend a lot of additional time fixing it up and makeing it flow seamlessly in a tool like AfterEffects or DavinciResolve.
The standard rate might be at 10%, but from my experience it still varieas a lot. I know of two B2B companies that deviate a lot:
* Company A has a product that practically sells itself - it has been joked that sellers are "just accepting orders" from prospective clients. At that company the commision rate is just 4% (used to be 8%). Most sellers are fine with that because it's easy money.
* Company B is not having such an easy time selling their product - at that company the commision rate is a whopping 30%.
While it's great to have some ballpark figures, you also need to realize that you may need to (or be able to) pay your sellers al lot more/less, depending on your specific situation.
There's a few companies already in this space, but all I'm aware of are of the "Contact our sales for pricing" type and require you to spend at least 100k annually. Many contracts go into millions - just for the apps & services, without media spend.
Isn't there a story about how a supercomputer used a macbook just to display a nice the logo on the door - and for nothig else?
ThinkGeek used to sell them, but that's not much help since I just discovered that TG no longer exists...
Another huge problem for these assistants is unusual band names. I have yet to successfully convince either Siri or Google Assistant to "play Einstürzende Neubauten".
Just yesterday I was reading an interview with the CEO of a German company which produces coronavirus tests about how they charge 2.5EUR for a single testing kit and that reasonable cost for processing it should be around 10EUR. Where does the 1000x multiplier come from?