im still a student so maybe im just overcomplicating things too early lol but lately ive been getting more annoyed w github actions the bigger my projects get. at first it was fine for basic CI and deployments but after adding more tests/workflows i started running into random queue delays, flaky runner stuff, weird marketplace actions breaking, all the usual things people complain about

the annoying part is i dont even know if this is a “real” problem yet or just me going too deep into tooling too early. some people around me are basically like “dude ur still a student just push code and stop caring about infra” which honestly might be valid lol. but ive still been experimenting w different setups lately because i wanted to understand the ecosystem better before internships/jobs. tried some self hosted runner setups, depot for a bit, and recently started messing around w tenki for some workflows just to compare how different approaches feel in practice

ig pretty much if whether i should stick with it or ishould switch up. is this just a skill issue for the most part?

im still a student so maybe im just overcomplicating things too early lol but lately ive been getting more annoyed w github actions the bigger my projects get. at first it was fine for basic CI and deployments but after adding more tests/workflows i started running into random queue delays, flaky runner stuff, weird marketplace actions breaking, all the usual things people complain about

the annoying part is i dont even know if this is a “real” problem yet or just me going too deep into tooling too early. some people around me are basically like “dude ur still a student just push code and stop caring about infra” which honestly might be valid lol. but ive still been experimenting w different setups lately because i wanted to understand the ecosystem better before internships/jobs. tried some self hosted runner setups, depot for a bit, and recently started messing around w tenki for some workflows just to compare how different approaches feel in practice

ig pretty much if whether i should stick with it or ishould switch up. is this just a skill issue for the most part?

just something ive been noticing lately. feels like a lot of people still use github for repos obviously but not necessarily for the whole workflow anymore

i keep seeing people move random parts elsewhere over time. self hosted runners, different CI tools, separate review/deploy setups, stuff like gitlab, forgejo, tenki etc. not even because github is unusable, more like people slowly getting annoyed at little things stacking up

if u still keep everything inside github what made u stick w ith it? and if not what part did u move away first

started pretty simple for me originally. one workflow, couple test jobs, basic deploy stuff. then over time more actions got added, more conditions, secrets, caches, matrices, random fixes copied from stackoverflow at 2am lol. now every once in a while i open one of the workflow files and genuinely have to stop and remember why something exists in the first place. especially the parts generated by AI months ago that technically work so nobody wants to touch them anymore

the weird part is i dont even think github actions itself is bad. it’s more like the ecosystem around it slowly accumulates complexity in ways that are hard to notice until later. marketplace actions, self hosted runners, release tooling, permissions, all that stuff kinda piles up. been trying different approaches lately mostly because i got tired of workflows feeling harder to reason about than the actual app sometimes. still using github actions for most things but ive also been experimenting w tenki for some CI/review flows just to compare how different setups age over time. maybe this is just normal infra brainrot tho idk

for a while i thought the ideal workflow was basically “type prompt receive finished article” but hmmm after testing a bunch of tools i kinda moved in the opposite direction. fully generated drafts always end up needing so much cleanup that the process starts feeling disconnected from actual writing entirely

lately ive mostly been using ai in smaller ways instead. stuff like brainstorming angles, restructuring messy drafts, shortening repetitive sections, cleaning awkward phrasing etc. chatgpt is still useful for random ideation, grammarly catches quick mistakes, and ive been messing around with writeless ai and claude more because they fit into editing workflows better than some of the giant one click content tools

honestly i think the ai stuff that survives long term is probably gonna be the quieter tools that help people think and edit faster instead of trying to replace the whole process outright

been jumping between different writing tools lately and i swear after enough use u start recognizing the exact same rhythm underneath all of them lol. doesnt even matter if its for essays, blog posts, emails, whatever. at first the outputs feel different and then eventually everything starts sounding overly polished in the same specific way

ive been trying a mix of stuff recently like chatgpt, claude, quillbot, grammarly, writeless ai, even some of the more niche rewriting tools and honestly the biggest difference for me now is whether the tool preserves the original voice at all during editing. some tools technically make the writing “better” but it stops sounding like something id naturally say after a few paragraphs

maybe this is just happening because everyone reads so much ai assisted writing now idk but i feel like people are slowly picking up the same phrasing patterns without even noticing it anymore

hi everyone

over the last few months ive been working on a side project called writeless ai

the original idea honestly came from frustration with a lot of the current ai writing tools out there. i kept trying different setups for drafting, rewriting, cleanup, “humanizing” etc and even though some were useful, the workflow always ended up feeling weirdly disconnected from actual writing. either the output sounded overly polished and robotic or the tools pushed way too hard toward full automation

so i started building something that felt more collaborative during the drafting process instead of just another one click content generator

right now the platform is mainly focused on things like

• rewriting awkward or robotic sounding sections naturally
• helping clean up rough drafts without flattening the tone
• restructuring messy writing
• making ai assisted text feel less overprocessed
• fitting into existing workflows instead of replacing them completely

ive been testing it alongside stuff like chatgpt, grammarly, quillbot, claude, stealthwriter etc and one thing im trying to focus on is making the editing feel lighter and more natural instead of aggressively rewriting everything

still actively developing it and honestly trying to figure out what people actually want from ai writing tools long term beyond just “generate more text faster”

if anyone here uses ai regularly for writing id genuinely love feedback on what current tools still do badly or what feels missing right now

tried a lot of ai writing tools over the last few months because i wanted something that actually fit into my workflow instead of just looking impressive in demos. most tools felt amazing for like 2 days and then i completely stopped opening them lol

1. chatgpt

still the most useful overall for brainstorming and rough outlining

what it does well

• quick idea expansion
• good for restructuring thoughts
• easy to use for random tasks

downside

• longer outputs start sounding repetitive fast

2. claude

probably the best for nuanced rewrites and tone

what it does well

• more natural sounding prose
• better at longer context
• good for detailed edits

downside

• can overexplain things a lot

3. grammarly

honestly still underrated for fast cleanup

what it does well

• catches awkward phrasing quickly
• easy workflow
• good for polishing drafts

downside

• sometimes pushes everything toward corporate sounding text

4. writeless ai

ended up using this more than i expected

what it does well

• keeps the original tone more intact
• better for editing rough drafts
• doesnt overprocess every sentence

downside

• still needs manual cleanup sometimes on longer pieces

final thoughts

hmmm biggest thing i noticed is the tools that survive long term usually arent the flashiest ones. its mostly the stuff that quietly fits into your normal workflow without trying to replace the whole process

i swear every month theres another ai tool that people say is gonna completely change how they work and then 2 weeks later nobody talks about it anymore lol. ive tried a ton of them at this point and most ended up feeling more fun to test than useful long term. some were good at generating fast output but weirdly hard to work with once i needed consistency or wanted stuff to actually sound like me instead of generic ai writing

lately ive been paying more attention to which tools i naturally keep opening instead of which ones have the craziest demos. ive still been using things like chatgpt and notion pretty regularly, and writeless ai has stayed in rotation mostly because it works better during actual drafting instead of feeling like a separate content machine. i think thats probably the biggest thing for me now, whether a tool actually fits into how u already work instead of forcing a whole new process around it

ive been trying a bunch of ai writing tools over the last few months and one thing thats been bugging me is how different they feel once u actually try building a real workflow around them. like at first its impressive seeing a full article or script appear instantly, but after a while i started noticing i was spending more time rewriting weird phrasing and fixing structure than actually saving time. a lot of them seem optimized around generating as much text as possible instead of helping with the messy parts of writing where youre reorganizing ideas and figuring things out mid draft

right now ive mostly been leaning toward simpler setups instead of full automation. usually ill draft something rough myself then use tools for cleanup or restructuring if i get stuck. ive tried jasper, copy ai, chatgpt, stuff like that, and lately writeless ai has probably fit best into my workflow just because it feels less aggressive about taking over the whole process. hmmm maybe other people are using these tools differently tho idk

Was working on a section about sanctions on russia for a paper and had this one paragraph where i was arguing that the short-term impact looked strong but the long-term effects were mixed. nothing crazy, just a pretty standard take. out of curiosity i ran that exact paragraph through a humanizer to see how it would sound. first pass came back more confident than what i wrote, like it turned the “mixed results” part into something closer to “largely effective.” second one did the opposite and made it sound way more skeptical, almost like the policy didn’t work at all

The weird part is if you skim both versions they look fine. grammar is clean, flow is better, nothing obviously wrong. but the position shifts just enough that if you handed them in, they’d read like two different arguments. i ended up ditching both and just editing the original myself. later tried drafting similar sections with writeless ai and it felt more stable in that sense. still needed edits, but it didn’t nudge the argument in a different direction every time i rephrased something. Made me a bit more cautious about running full paragraphs through humanizers now. it’s not just tone getting adjusted, sometimes the actual stance drifts without you noticing until you compare side by side