I am making faceless youtube videos and treating it as a long-term bet, even if some people say it's oversaturated.
Stack:GPT(scripts + image generation) + minimax (audio) + Capcut(edit)
Output still feels low quality, should I improve visual quality first, or is script/hook the real priority?
what would you improve first? And what 's your faceless video stack?
I used ai to do a research pass on ai coding tools, mostly to gather material for an article, and the output was more interesting than I expected.
It also made me notice how "algorithmic gravity" shapes what feels "true" online: in my feed and circles, Claude Code (and maybe Codex) feel like the default answers, but in the synthesized report, Github Copilot came out on top, and tools like replit and bolt ranked ahead of Claude Code.
The mismatch is genuinely fascinating to me.
[ Removed by Reddit on account of violating the content policy. ]
I launched my first digital product last week, and so far sales have been underwhelming. I have no experience or background in this area. I asked ai for ideas, and it gave me a content calendar built around: improve SEO and distribute on TikTok and Pinterest.
What I'm trying to understand: For people who've actually done this, is that advice generally effective for a new digital product, or is it more of a checklist that only works with specific conditions (audience, niche, offer, price)?
If you have experience with any of this, I'd love your take on:
- whether SEO + tiktok + pinterest is a realistic path in the early days, or if I should prioritize something else first.
- what you'd do this week if you were in my shoes
- common mistakes you see first-time creators make right after launch
Thanks in advance, really appreciate any practical guidance.
I‘m freelance /small-business side now, so I've had to get way more practical about AI spend than when a company could cover tools.
My rule so far: Keep costs as low as possible while still shipping market-comparable quality. I moved my paid chat sub from Claude to ChatGPT (Claude felt weaker for me lately; GPT’s image gen is strong), and I still pay for Cursor for local-file workflows and multi-model access. Canva and CapCut I use free—good enough for now.
I'd love to how others think about this:
- what ai tools do you pay for
- what did you drop first when budgets tightened?
- any free combo that replaced a paid tool for you.
This year I left an AI startup and went freelance while building my own small business.
The tradeoff is straightforward: my income is lower than before, but the goal is to get the business working end-to-end and make it stable — so I have to be cautious about AI tool budgets.
At my last job, my stack was basically “whatever works”: ChatGPT Plus, Claude Max, Gemini, Cursor, Midjourney, and Lenny’s product pass — partly because the company could cover some of it.
Now that everything hits my own card, I’ve had to get a lot more practical.
Chatbots + image models
Since last month, Claude has felt like a noticeable step down in quality for me, while ChatGPT’s newer image generation is genuinely strong. So I switched my paid chat subscription from Claude to ChatGPT.
I also tried Nano Banana; Banana 2 is paid as well. Beyond pricing, single generations were fine, but multi-image consistency for the same character wasn’t stable enough for what I need. I’m very used to Midjourney’s aesthetic, but it’s expensive — for my use case right now, ChatGPT is the better value.
Other than GPT, I pay for Cursor too (local file handling and execution is often more convenient.)
I’m still on Cursor because the paid plan includes API access across major models (Claude, Gemini, GPT) plus Cursor’s own models. The UI has changed recently, and there’s more of a chatbot-style workflow now — which fits how I work.
Side note: I used to build my own tools (I’m not a classically trained engineer — more Claude Code + Cursor “builder mode”). But turning a personal tool into something with real users and real revenue is a long runway, and I still need to level up on coding. My focus for the past year has been content — social, digital products, AI product growth — so I dropped Claude Max as part of tightening spend.
Everything else: free combos + “boring” traditional tools
For a lot of other needs, I stack free products and replace AI with traditional tools where it’s good enough.
For digital products + faceless social video, my current combo is ChatGPT, CapCut, MiniMax, and Canva.
Mini-rant: I used to subscribe to ElevenLabs, and I still have credits sitting in the account — but I can’t actually use them unless I resubscribe, which feels pretty user-hostile.
I wanted to do heavier AI video. At my old company I’d top up / expense that kind of thing, but AI video is still expensive on a freelancer budget, and a lot of products still feel like “gacha” (you keep rolling until you get a usable take). So I landed on faceless video instead. After skimming tutorials, it’s surprisingly simple in execution — the real bottleneck is the script.
My rule while I’m not profitable yet
Keep costs as low as possible while still shipping market-comparable quality. So far, it’s working in practice.
That’s my current setup — what would you recommend instead (or what’s your “freelancer survival” stack)?
For me it's mostly small, personal needs - I like building tiny things that solve my own use cases(nothing fancy).
Could be really small (a morning checklist, workout logging, language practice) or something a bit bigger.
What about you?