William Roush (@StrangeWill) profile picture

Founder, Investor, Engineer @RoushTech | Building @RaptifyApp (and a ton others). Software, virtualization, storage, networking, VOIP & more.

Soddy Daisy, TN
Joined October 2011
Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)Photo by William Roush (@StrangeWill)
William Roush retweeted
118
4,912
155
62,320
1,009,860
William Roush retweeted
I'm really near the end of my rope to the verge of crashing out about AI bullshit. Do I use it? Yes. Do I have to choke the shit out of it over most non-trivial things? Yes. Fable? *YES* How the fuck are you people just letting this thing go on autopilot? Then you have dipshits being like "oh it can do 100% of my job" This is how >80% of businesses are finding LLMs are shipping bugs constantly. GFDI Stop flagging you're massively shit at your job. We've been through 3 years of some of the worst fucking software development I've seen in my life, bugs, slop, garbage, noise, people not giving a fuck, rather Reddit and YouTube while Claude ships shit. You like it because you're fucking lazy and you want to fuck around on your phone instead of work, especially while collecting a check. The rest of us are dealing with the worst fucking software we've used in our fucking lives. If it's so fucking great, then produce better fucking experience for the rest of us. If you can't do that, then please shut the fuck up.
9
3
39
2,400
William Roush retweeted
“there's approximately 0 value in knowing how to code” *checks profile* AI startup Every time
the "software engineering" i was doing in 2022 has been fully killed by fable for a minute i thought it wasn't that different to opus/codex, but once you get fable managing sub-agents, it becomes clear that it's 10x better i used to be half joking when i said i wasted years of my life learning programming languages by hand, but now that's just 100% true there's approximately 0 value in knowing how to code if you want a job building things today, you better know how to actually build product end to end. don't expect that you can just be a CRUD monkey anymore you need to have your own ideas & vision, be capable of sending outbound & taking feedback, iterating, marketing, improving, deciding when to stop loss and pivot, then running through that loop on repeat the only value left in us meatbags is our ability to be accountable & own things end to end
54
83
3
2,440
54,737
Continue seeing the metrics you know and love from Mantle. Self-host if you like. Sign up for early access.
11
2
2
38
4,455
William Roush retweeted
Affilitrak just hit 200 reviews!
5
1
55
1,591
William Roush retweeted
Affilitrak for Shopify Apps is live! You can easily import your data from Mantle and test it while keeping your Mantle affiliate program running. Try it in beta for free at affilitrak.com/for-apps
To the Shopify app devs who used Mantle affiliate: Affilitrak is launching "Affilitrak for Shopify Apps" More details below 👇
7
5
2
36
4,718
William Roush retweeted
Replying to @LukasHozda
Also: AI means apps have a much higher bar now, they need to do more, if you're slapping them out in a weekend then someone is spending 3 weekends and having a much more polished product. Someone else is putting in 3 weeks and has an app 10x better than yours. The ending math hasn't changed, but the bar for what is *sellable* got a lot higher.
1
1
2
46
William Roush retweeted
Replying to @ThePrimeagen
To be fair, a lot of engineers could barely fucking code before AI, and IMO: if you can't do both you're NGMI especially with how expensive it's getting. You're going to have to determine when it is and isn't worth it, and your skill to get things done efficiently, which includes understanding good code-based approaches, is going to be important.
1
1
344
William Roush retweeted
But also: people need to stop vibe-coding custom software instead of using things like page builders, y'all giving me an ulcer on how many marketing landing pages are react apps instead of just some CMS.
1
1
32
.NET is the best developer platform to build your apps on right now. And it's not even close.
10
14
182
7,665
William Roush retweeted
How did Windows become such a dogshit OS?
786
105
92
2,803
216,099
William Roush retweeted
In today's 996 chest-thumping culture, don't forget about the quiet ones keeping your company together..
92
338
128
2,645
644,349
William Roush retweeted
Replying to @voroninp @kellabyte
Haven't read but I'm familiar with the concept -- I generally refer to shallow modules as "not abstraction but *misdirection*", I've never used code with them that didn't increase frustration while not really netting me anything. The problem is that ORMs by their very nature are *extremely* leaky and it's hard to compartmentalize them (again, in a way that projectables don't just do, but better). Abstracting for the sake of having an abstraction is a waste of time and overhead. Our team usually adheres to things like WET (Write Everything Twice) because writing abstractions up-front is massively error-prone (and a lot of the time I may not even need one). *Especially* with ORMs this gets troublesome, and is why I buy into the Testing Trophy, for anything that isn't a low-level library -- something with a lot of business logic and data slosh, I will *not* mock the DB as part of testing, we test against real DBs -- this usually ends up with a "fat" integration test layer but it means my tests are validating that the software works, not that my software builds database-impossible conditions that won't get caught until it's in front of the customer. In general a lot of .NET devs are "I worked at $bigCorp and we did this dumbass thing" or "I read a book 20 years ago in college and... [applies it incorrectly]" And our software suffers greatly from it.
2
1
9
688
William Roush retweeted
Replying to @kellabyte
I've had to argue with engineers to not wrap a repository pattern (EF Core) with another proprietary Repository pattern. This resulted in their team lead getting involved and allowing it. Our S-tier engineer has been spending days ripping it out because of all the fucking problems it causes (with no benefits over just using projectables) I love C# but man our market is full of trash engineers with trash in charge that'll approve it.
12
1
1
118
14,140
William Roush retweeted
Replying to @AntonMartyniuk
Repository patterns made sense with ADO.net and sprocs With EF Core it's already a repository pattern, so you're just wrapping a flexible repository in a less flexible repository you manually maintain It's madness, I've never seen it improve a code base, always makes it significantly worse.
1
3
268
William Roush retweeted
is it just me or AI coding models are hitting a plateau? building apps with it hasn’t really gotten that much better over the last year. feels pretty much the same tbh...
23
6
2
123
13,873
William Roush retweeted
Replying to @eepyeri
NGL, I've faked server version headers (or removed them) entirely because security scans are piss-brained. Seriously, I've had to talk to security firms about, yes, when you patch software they don't increment version numbers, this is how Linux has worked forever, this is why repos are "frozen" in time, this is what leads to a crazy amount of stability. Yes, it's 1.18.0-ubuntu9, it's patched this bug, fuck OFF. Then I just hide the version because I'm tired of them not understanding. Oh look, we're not longer vulnerable. JFC.
2
1
22
3,175
William Roush retweeted
Replying to @AdamAkhlaq
I uh... this depends on the product and the market. We've had products that $100/mo is a cakewalk, and have had products priced in the $1k/mo range easy. We've had products that sit nicely in $10/mo. Our current product we're building at @RoushTech sits nicely at $30-70, with *maybe* an enterprise feature set around $150/mo, you just have to be able to justify the cost/benefit ratio. I wouldn't even say "a cheap price means a cheap product" like others have mentioned here, sometimes you make it up in volume, sometimes the app is *good* and polished but just basic and doesn't justify a high cost in the time-save. The underlying strategy: Know your product, know your market, know your value, price accordingly.
1
14