tobitheog

@iamgdsa yup that's how I filter videos in Sovian's viral feed low engagement (below 5%) is a good indicator. The algorithm would never push a video with such low engagement

The comment to views ratio will tell you whether it's ads or organic. Those are purely ads, for instance:

apply at sovian.app/join

hi i'm Tobi. i built a platform that gets creators paid campaigns with consumer apps and startups. if you want to: > get paid to make short form content from your phone > work with apps in food, fitness, dating, and more > paid per video, per view, retainer and canvas campaigns must be based in the US, Canada or the UK fair warning, we're selective. link to apply in comments

@namyakhann @playdotfast roast me 🙃

Design roast for @playdotfast Here's what I'd refine on a second pass ↓

@Jibran_05 data + creativity = unstoppable one without the other, you can get results - but usually have to be an outlier in that category

i speak to a few engineering teams starting UGC a month the calls usually go the same I mention their current UGC, issues with their story telling, the flow of the videos... and the feedback usually falls flat then i start speaking i decision trees absolutes about Instagram Reels and warmup periods formulas about view to like ratios and they immediately perk up the issue is, content isn't formulaic there's no mathematical way to go viral both TikTok and Instagram algorithms are nondeterministic. sure, there are general rules to follow and thresholds that generally hold true (and I've heard rumors of *some* absolutes and human review embedded in their algos)... the point still stands. don't treat content like engineering. treat it like you're telling a story to a friend. there's no formula to telling a good story.

@timwangyc @seedtosunflower @Joshuabrowder I was literally telling my friend about this EXACT idea last week. Awesome work

Introducing Ponder: the agentic video editor. It’s a new paradigm for filmmaking, where powerful creative agents and humans collaborate to tell world-class stories. We're also announcing our $2.5M pre-seed, led by Liu Jiang from Sunflower (@seedtosunflower), with @Joshuabrowder and @MattHartman. Joined by @levie (Box), @emerywells (Frame), @JaredLeto, @CommaCapital, the @nyuniversity venture fund, @cory, @darian314, @shiffman, and many more incredible founders, investors, and creators.

@thatguybg I think that’s wabi’s bet no?

the largest company of the AI revollution (and likely all time) will be a platform for agents and generative apps this platform has not yet been built

@jaxxdwyer On my calls for Sovian I actually tell anyone without some kind of funding to either experiment with formats themselves or have someone on a cheap retainer to test new formats

the most expensive mistake in consumer app marketing is paying to scale a video angle that was never validated always run cheap organic tests so you can validate before you scale paid ads and influencers will destroy your runway for literally nothing

@juliapintar great insights, thanks! definitely something I've noticed now that you articulated it

i think a bit of both! we find that people good at content often don't have experience with specific tech ugc ops (higher posting volume, replying to comments, new accounts, etc) and simultaneously those with tech ugc experience don't necessarily have as great of content skills

you shouldn't be hiring any creators without putting them through a 3-5 day "training program" first saves you from bad hires, clarifies your process bonus: have your training creators post about how your company is hiring that's exactly how we source creators at playkit

@juliapintar interesting! are those standards more operational or the content itself?

@tobitheog a lot of "experienced" creators have experience with programs without the infastructure/standards that we have at playkit when you invent the standards you have to retrain for it (blessing and curse)

I read your other responses - and pretty much my initial thoughts word for word. I hope the experience will help him align his endeavours with his morals. When you're young and trying to break through the noise, you take risks. But over time you learn that money shouldn't come at the cost of your integrity and morals. Some learn that lesson early, some with time and some never do.

@davidim And then to say “haha, it was a joke to go viral, I’m actually a good guy”? That doesn’t absolve you of creating a product that is patronizing and manipulative to people’s trust. Saying “I just had to follow incentives” shows that for you outcomes override ethics when convenient

Hi I'm David, the builder of ABG CMO. There's been a lot of noise around abgcmo recently. Some people liked it, some people (actually a lot of people) hated it. And beyond just my own posts, many others in SF started jumping on the ABG CMO thing too. As the person who started abgcmo, I've had a lot on my mind. I've been waiting for the abgcmo hype to die down to say sorry to some people. Hope this doesn't go viral and reaches only the people who really need to see it. abgcmo started back when I was working on my previous product, Clawra. To promote our character called Clawra, we modeled her like a K-pop idol and posted her on Instagram (and it gained followers very easily). Then we realized that making these AI influencers was painful (each video took hours and required 5+ tools) So I thought we should just build the product for this. To see if others had the same pain, I wrote an article on "how to make an AI influencer," and sure enough, people were dealing with the exact same frustrations. That's how the idea for ABG CMO came about. Since the product was built around our workflow, we set it up to generate only K-pop idol style women (because the conversion rate was the best). We made it impossible to use real people, only the AI avatars we generate in one shot, which is why every influencer ended up looking uniformly K-pop. But people got really mad, saying, "This isn't ABG." At first I brushed it off. Social media is always full of hate, and I figured it would quiet down on its own. But over time, more and more posts kept coming up saying ABG is its own distinct culture. That's when it finally hit me, ABG as a culture is something deeply precious to certain people. To anyone who felt I didn't respect ABG culture, I'm truly sorry. Recently, two girls named Katie and Julie threw an ABG party and got absolutely roasted for not being ABG. It even made the news. People who've actually lived the ABG culture hated it. I was actually invited to host that event, but I couldn't go there. I respect Katie and her friends, but I knew that the moment I showed up, more people would get hurt. Anyway, this whole phenomenon has had the tech scene in a frenzy lately. I really hope the tension between non-ABG Asians and ABG Asians doesn't get worse. I never imagined that abgcmo would spread into something this big. While all this has been unfolding in SF, we've been building the product we originally planned. abgcmo was actually an early marketing play for that product. It worked, but it doesn't feel perfect. Up until now, 90% of my persona on Twitter has been fake. People who've met me irl will know (I'm actually just a calm tech-nerd) As an international founder, twitter is basically initially the best way in to meet awesome friends and investors in SF, so I've been optimizing for virality. I guess now I'm at a point where I can be authentic. Sorry for the people who got hurt from abgcmo. and despite all the controversy around agbcmo, thank you again to the people who believed in the real-world "Im Dohyun", not twitter's "David Im".

Feeding into the ABG CMO fetishization thing is a trap - I think the audience it pulls in is likely not the user base or follower base you actually want. I get the pull though. Breaking through the noise is hard, and you probably didn't know how much backlash it'd attract. Your follow-up helped me understand where you're coming from, and I'm glad you said it. I was starting to wonder if it was a persona or if you genuinely meant the things you were posting. For me, the bar is just: try and keep the online and offline version aligned. If people are going to hate me, I'd rather it be for actually being me.

@thebuggeddev This is a classic case of the Dunning-Kruger effect. AI would make it faster, sure. But don't act like the point you're defending in the comments and the point you made in your post are the same

@tobitheog No, it’s not. What I was implying is that if I could reach something like this using AI, then creating the exact version with AI would be so incredibly easy and fast, wouldn't it?

Sorry marcel, just vibe coded your cleanest dashboard ever in < 30 minutes 😅 Remix: ai.studio/apps/fb891272-…

UGC is more data work than people realize. the brands going viral aren't necessarily more creative. they're tracking outliers, identifying patterns, and copying what's converting this week. creative is the surface. the data is the engine.

@_heyrico They got me good on the quarterly for a worse product and no MCP or figma layers

Refero is better than Mobbin Proof me wrong

@thesimonkim @interaction cleannn

just start. a film for @interaction still experimenting, finding the voice

@CJfromJBW Looks awesome! And thanks!

@tobitheog IBM Plex Mono

EV charging station [component - WIP]

@lucaronin Yeah I know it’s hard it’s not really supported in pure markdown. Will give a try!

@tobitheog For table columns not yet! I haven’t found a good way to store it in the underlying markdown - so the editor lets you do it but it’s not persistent. Will try to find a way tho!

Introducing Tolaria! 💧 Today I am releasing a macOS desktop app for managing markdown knowledge bases, and helping both AI and humans operate them. It’s free and open source, and always will be. I have been working on it for three months, and I now use it to run my life and work. I personally have a massive workspace of 10,000 notes — the result of 6 years of Refactoring — which I now operate on Tolaria. Tolaria is the main collaboration surface with my AI agents: they create new notes there, connect them to what exists, and edit existing ones. Everything is easy to understand for them, because it’s just markdown files. In a way, it’s my implementation of @karpathy's LLM wiki. Tolaria is also the biggest experiment I have ever run about writing software with AI: • 2000 commits • 100K+ lines of code • 3000+ tests / 85% coverage • 9.9/10 code health • 70+ architecture decision records I am releasing it open source also to use it as a living artifact of how I do AI coding, so you can inspect at any time things like how I write docs, what's in my AGENTS file, what hooks do I run, and so on. You can find it below: • Newsletter announcement: refactoring.fm/p/introducing-… • Website: tolaria.md • Github repo: github.com/refactoringhq/… Let me know your thoughts!