Whenever I browse product launches, I somehow subconsciously judge not only the product itself and its quality, but also the quality that is reflected in the effort the makers put into preparing it.

It may sound insignificant, but in my case, these things also make a significant difference:

• Icon GIF at the launch it enlivens the overall impression and is dynamic

• Quality graphics and video

• First, a properly filled-out comment

• Photos in the makers' profiles (it's less trustworthy for me when there's only the letter "J" or something similar)

• Whether any of my contacts or acquaintances on the platform reacted to the launch

When you build many products based only on a feeling, you eventually realise that the majority of them won t work.

Sometimes it s because the distribution wasn t good enough. Other times, the product itself wasn t good, or people simply didn t need it.

Formally, I studied marketing, but honestly, that stuff from textbooks didn't help me at all. :D (sounds like wasting 5 years of my life, nwm)

The best marketing things I have learned weren't from courses, but through:
own projects

calls with someone better than me

I often see the media sharing articles about layoffs due to AI, how junior programmer positions are less in demand, how there is also a decreased interest in copywriters and graphic designers, etc.

About 2 weeks ago, Teammates launched a tool (AI HR-ist), and right now I came across a post from a local marketer who shared interesting data about Ask AI (an internal AI/chatbot system), which today handles almost 94% of all routine HR requests, such as:

• vacation requests

• onboarding new employees

• payroll information and attendance records

• benefit selection and answers to basic employment questions

Results of AI implementation at IBM

• 94% of the HR agenda is automated

• Payroll, vacation, administration even terminations have been automated

• $3.5 billion saved

• 40% drop in HR costs

IBM also claims that employees are happier. The HR department s internal NPS score increased from -35 to +74 after the implementation of AskHR (source: HR Asia). 6% of questions are still directed at people AI has not yet completely replaced complex or emotionally sensitive situations.

I was dealing with an issue where automated scripts kept trying to "optimize" small stuff instead of just finishing the main task. You ask for a simple feature update, and suddenly it's changing file structures and refactoring background logic nobody asked for.

It feels like the more context you give them, the easier it is for them to get distracted by non-essential tasks.

For people building with cursor or agent setups daily, how do you lock down scope so it stays focused on a single file or task?

I spent hours last week re-checking an agent's work because it skipped core steps from a spec I explicitly approved.

Having to manually audit thirty different file changes completely defeats the point of automating the build.

Now I write a separate reference file directly from existing code and force the runner to compare against it before marked complete.

How do you handle automated coding tools going off-script without spending all your time reviewing PRs?