Roman

10.8K posts

Roman banner
Roman

Roman

@romxdev

Software engineer with impostor syndrome figuring out the indie SaaS game

Remote Katılım Mayıs 2021
1.5K Takip Edilen1.7K Takipçiler
Roman
Roman@romxdev·
@darekgusto I'd definitely enjoy seeing more about what you're building 😉
English
1
0
1
12
Darek Gusto
Darek Gusto@darekgusto·
Taking some time to rethink my approach. Well, that and trying to not waste the last two days of my Cursor billing cycle. I might be tweeting less. Or more, or some totally random stuff. Or something actually about my building adventure. Or nothing at all. Thanks for sticking with me.
English
4
0
7
147
Roman
Roman@romxdev·
@darekgusto Thanks 😉 He's slowly building his own fan club
English
0
0
1
8
Roman
Roman@romxdev·
Your career will survive a 10-minute walk away from the screen. Your back might not. Take care of your body before it forces you to.
Roman tweet media
English
2
0
2
37
Kotlin by JetBrains
🔗 Your JetBrains IDE experience can now be showcased on your LinkedIn profile. The free LinkedIn Connected Apps plugin generates a proficiency level based on how you actually use your IDE. It works in the background and updates your level automatically as you explore more IDE features. Check your current level from the status bar and securely connect your profile through LinkedIn OAuth when you’re ready to share. Give it a try 👇 kotl.in/zsxnxi
Kotlin by JetBrains tweet media
English
2
3
36
4.9K
trash
trash@trashh_dev·
i have tokens to burn and no ideas
English
74
12
198
12.2K
Roman
Roman@romxdev·
@Zhuinden You definitely need an academic degree for something like that 😂
English
0
0
0
7
Gabor Varadi
Gabor Varadi@Zhuinden·
@romxdev I thought "soft skills in tech" was more about knowing how to assign fibonacci numbers (or t-shirt sizes) to tasks of unknown duration or complexity
English
2
0
0
147
Roman
Roman@romxdev·
Soft skills is a misleading term in tech. Being able to explain a complex technical problem in simple plain English is actually a rare hard skill
English
1
0
1
173
Roman
Roman@romxdev·
@RyanEls4 That's how we used to code in university 🥲
English
0
0
1
17
Ryan Els
Ryan Els@RyanEls4·
The real legends 💪😁
Ryan Els tweet media
English
22
3
52
980
Roman
Roman@romxdev·
@ActualJC That works well for explaining what the code does. But it doesn't explain why the team made certain decisions, what was tried before, or what to avoid repeating.
English
1
0
4
108
James
James@ActualJC·
@romxdev “Claude, please document this entire codebase, then produce a high level plan for reigning in the complexity and another plan for a testing strategy”. Seriously, in 2026 documentation is solved.
English
1
0
4
133
Roman
Roman@romxdev·
Today I realized that most dangerous developer on any team is not the low performer. It's that ultra-fast developer who pumps out thousands of lines of clever code with zero documentation and then leaves the company.
English
40
7
123
5.6K
Roman
Roman@romxdev·
@darekgusto It depends on the content 😉 Sometimes I understand that there are no ideas for image
English
0
0
1
7
Darek Gusto
Darek Gusto@darekgusto·
@romxdev Then a different angle... how do you decide which post gets a cat pic?
English
1
0
1
17
Darek Gusto
Darek Gusto@darekgusto·
A vanity check! When you write a new tweet on your timeline... do you have some performance goals in mind, that make it a hit or a miss for you? I mean comments, likes, impressions and other metrics. --- For me, if it's just a casual post, I'm satisfied if the post gets 10+ comments and 10+ likes, with 400+ impressions after 24h. If it's a post I've spent more time on and think it's got more substance, I aim for 15+ comments and 15+ likes, with 700+ impressions. I consider 20+ with 1k impressions a rare banger but I try to aim for it more often. But for now, I often hit that only because something else happened, like it was a tweet on timely topic with the right keywords, or one of the comments blew up rather than my post itself. No reposts or bookmarks, since it's not the type of content I usually write. Record for own post? 13k impressions, but only because steinpete blessed it with a comment.
English
16
1
17
577
chrisreedbates
chrisreedbates@chrisreedbates·
@romxdev Thats like 90% of X that switch models as soon as a new one comes out from a competitor.
English
1
0
1
278
Roman
Roman@romxdev·
@we_arent_here I liked it because I think you're right that AI can generate documentation from code. I just don't think that solves the problem I was talking about. It's great at reading code. But it's much worse at explaining decisions that only existed in someone's head.
English
1
0
0
8
Religion is poison
Religion is poison@we_arent_here·
@romxdev what the fuck why did you like this tweet, did you even read it? or is it your engagement farm just automated?
English
1
0
0
5
drstan 🇨🇭
drstan 🇨🇭@drstan2023·
@romxdev the team is already a bad team if you have developers at all. you should have vibe coders.
English
1
0
1
83
Roman
Roman@romxdev·
@FnMiddle I was in this situation, when one developer decided to write a separate component in Scala instead of using the familiar Java 🤷‍♂️
English
0
0
0
112
Kevin Golding
Kevin Golding@FnMiddle·
@romxdev Especially when writing it in a language that maybe 3 people on the whole planet know, eg. unicon.
English
1
0
1
164
Ignazio De Santis
Ignazio De Santis@ignaziodes·
@romxdev the person isn't really the risk, the incentive system is. that developer gets praised for the lines shipped and nobody's job is to measure the debt until eight months after they've left and someone else has to pay for it
English
1
0
11
263
Roman
Roman@romxdev·
@AdamLGRing Not in my case tbh 😅 As a result, we have a lot of people in wrinkled T-shirts on the call
English
0
0
1
8
Roman
Roman@romxdev·
The biggest lie of remote work is that we turn on cameras during meetings to connect better as a team
English
2
0
2
779
Roman
Roman@romxdev·
At this rate, tech companies will soon start paying our salaries in tokens instead of real money.
English
4
1
7
209
Dmitrii Kovanikov
Dmitrii Kovanikov@ChShersh·
Holy shit, I just discovered a much better interview approach than LeetCode
English
58
5
813
139.3K