After experiencing many bugs and UX oddities with both of those, I went back to ScreenStudio.

ScreenStudio is reliable and produces the best results for my use case (educational content and client updates)

Useful for compositing on top of explainer videos, or just as a blackboard app.

Comments URL: https://news.ycombinator.com/item?id=47324746

Points: 2

# Comments: 0

Vibe coding does none of the above

Comments URL: https://news.ycombinator.com/item?id=47036144

Points: 1

# Comments: 0

If you are feeling over reliant on these tools then I quickfix that's worked me is to have real conversations with real people. Organise a coffee date if you must.

I'm a software engineer in games and big tech for a combined 15 years. Coding is my biggest passion and I'll still do it when I retire.

However, a large part of me wants to get into an entirely different industry.

I've identified the reason to be how intellectually unstimulating coding has become (due to vibe coding).

The AI tools are incredible - but in terms of learning and self development, vibe coding as an education form lacks depth and the classic sense of mastery. I feel like I've learned 90% of everything there is to know about vibe coding. The last 10% is just marginal gains.

I live for the thrilling moments in software development - I was in the industry early enough to roll out software on our own server racks. As a game dev, I regularly implemented algorithms (such as path finding, physics integrations and rendering logic). I also got to work with Mixed Reality headsets and help define UX paradigms.

I feel like there's now a stigma around DIY coding - one should simply use an off-the-shelf solution or AI.

I get so much pleasure in doing deep work, but nowadays, any feature that takes more than 2 days to implement gets eyebrows raised.

Am I overthinking it?

Comments URL: https://news.ycombinator.com/item?id=46559728

Points: 7

# Comments: 9

Feels appropriate: What if we made advertising illegal?

I agree with CLI feeling more deliberate. It could be the lack of "accept" gates (that you get with Cursor, Copilot etc).

When you say "switched to Claude Code entirely" do you still make manual edits in a text editor is CC doing 100% of the coding?

Also, you mentioned getting good at prompting - is there anything you do that's not already well documented?

Points: 3

# Comments: 7

Source: I worked there.

My client, WhipSnap, wants to serve the niche of 3D car scanning - applications range from enabling car showrooms to conduct virtual sales and marketing to letting petrol-heads scan and share their own cars, with annotations in 3D.

I designed and developed an MVP for them. I also visited high-end car showrooms with my client to scan cars. I was also the "tech guy" in calls with investors (I rarely spoke, TBH).

Here's everything that led to this moment:

[-2]. Pitched to a different client.

[-1]. Got rejected.

[ 0]. Lesson learned: needed to dramatically increase the quality of my proposal.

[ 1]. Heard from a former colleague about another project

[ 2]. We setup the initial call for intros

[ 3]. Before the call, I worked the weekend to research their niche and the technologies they could use.

[ 4]. During the intro call I impressed them with my research, my understanding of their niche and suggestions on how we'd build their app idea.

[ 5]. They signed.

[ 6]. During the project, I sent weekly video updates via Youtube (unlisted). I'm a fan of async comms.

[ 7]. We spoke regularly, via a WhatsApp group chat!

[ 8]. Within 4 weeks we had a useable MVP. Real customers could scan their cars!

[ 9]. Then some chaos: I personally helped scan a $4.5M Koenigsegg.. and the scan failed

[10]. More chaos: Runpod (the GPU cloud provider) became unreliable. I ended up buying an RTX 5090 to run our 3D pipeline on.

[11]. The weekly videos and regular comms kept my client informed and calm. Together, we worked around issues.

[12]. I delivered and hosted the solution. Did some extra bits to increase the quality of car scans (like manual cleanup).

[13]. We onboard early adopters.

[14]. Investors understand and are excited by the solution.

[15]. Investment money arrives in my client's bank account.

I appreciate some of the above doesn't sound scalable. I also appreciate that it sounds unconventional (like using WhatsApp group chat with clients). But they worked for me.

Thanks for reading!

p.s, my company is https://llume.co, you can find a more in-depth case study there - see "Carvatar"

Comments URL: https://news.ycombinator.com/item?id=44073579

Points: 1

# Comments: 0