We turned AI wait time into a tradeable market. Developers earn, advertisers reach.

I look at a lot of two-sided marketplaces, and the ones that survive seed one side with something that works single-player. Your developer side has that with passive earnings, but the advertiser pitch lives or dies on proof that wait-time attention converts, so publish click and conversion benchmarks from your first campaigns as soon as you have them. One anchor advertiser with a public case study will do more than any feature you ship this year.

Love the creative approach to a problem we all secretly deal with. Monetizing the "spinner stare" is a clever angle.

The biggest challenge I see for two-sided marketplaces like this is avoiding a shrinking-inventory problem as developers get better with agents. A lot of heavy users are training themselves to parallelize their work—kick off a prompt in one terminal, and immediately switch to another task or workspace rather than waiting around.

A couple of quick thoughts/questions:

• The Privacy Hook: Since devs are naturally skeptical, have you considered open-sourcing just the signal-detection layer? Showing exactly how you detect the AI is busy without reading prompts or code would build instant trust.

• Targeting for Advertisers: Reaching "developers" as a broad bucket can be hard to price effectively. Will advertisers be able to target specific niches (e.g., frontend vs. devops, specific frameworks, or MCP server users)?

I’d definitely be curious to see how engagement and click-through rates look once you publish early benchmarks. Wishing you the best of luck with the launch!

The idea makes sense if the ads stay highly relevant. AI wait time is one of the few moments where I'm not actively typing. I'd be curious to see how you keep the quality high over time.

As a developer: the honest reason I'd keep this installed isn't the earnings, it's whether the card ever feels like noise. Passive income from attention I'm giving away anyway is a nice-to-have, not a reason to install. The real hook is if the discovery is actually good. If the card surfaces an MCP server or a tool I genuinely didn't know I needed, a few times a week, then I'm not watching ads during waits, I'm getting a curated feed of dev tools at the exact moment my brain is idle and open to it. Although Im unsure I want to further get distracted during a deep focus session.

On the parallel-agents point others raised, I hear it, but I don't fully buy that it kills you. I'll be honest, on a bad day I've got 6 or 7 sessions running at once, so in theory I have zero wait time. But even then there's this specific micro-moment right after I fire one off and before I've loaded the next thing into my head. That gap is tiny and it repeats dozens of times a day. Short and receptive might actually be better inventory than long and distracted, so I'd measure it before assuming the trend buries you.

I'd genuinely pay to reach developers in this moment, but three things decide whether it's a no-brainer or a pass:

First, targeting. Reaching "developers" is too broad to price. If I can somehow target a specific ICP it would be way more valuable.

To your direct question, yes the "earn during AI waits" idea clicks, but I think leading with earn undersells it. The sharper pitch to devs is that the best tools find you exactly when you're idle, and the earning is the fairness mechanic underneath. For advertisers, lead with attention given by opted-in devs.

Clever use of dead time. One thing I'd push on that hasn't been said: the premise assumes wait time is idle attention, but the better a developer gets with agents, the less that's true. The workflow everyone's moving toward is parallelizing, kick off one agent, switch to another task, never watch the spinner. So your best-fit users (heavy agent users) are exactly the ones training themselves to have zero wait-attention. That's a shrinking-inventory problem baked into the trend line. Worth thinking about whether you're monetizing a behavior that's on its way out.

The privacy claim is your make-or-break though. "Watches your AI workflow but never reads code or prompts" is a hard sell to devs on faith. Open-sourcing just the signal-detection layer would kill the skepticism faster than any explanation.

I actually like that you're treating the wait as the product instead of trying to interrupt the workflow. My biggest question would be whether developers still perceive that time as "available" attention once they're juggling multiple agents in parallel.

Interesting concept! Turning AI waiting time into a marketplace is a creative way to monetize otherwise idle moments while giving advertisers an opportunity to reach a highly targeted audience. I'm curious to see how you'll balance user experience with ad relevance so the waiting screen feels valuable rather than distracting. It would also be great to share some early metrics such as engagement, click-through rates, or developer earnings as the product grows. Wishing you success with the launch and looking forward to seeing how Backspin evolves.

This is a really interesting idea, and seems like an excellent way to make use of the idle time that agents currently create.

I'm curious, though: as I've used agents more extensively, I've focused a lot on parallelizing the work I'm doing -- switching among workspaces in cmux, for example, so that I'm not left staring at the AI while it works, but am instead working on a different workstream while the original agent churns away unseen. The fact that you're paying folks for their attention, combined with only presenting a "single relevant card", should help mitigate this -- it aligns the incentives nicely.

I'll install it on my home machine tonight and take it for a spin. Looking forward to seeing what you've build here!

interesting idea, the “idle time has value” angle is actually pretty clever

but honestly my first reaction is that devs might not like anything showing up during that moment, even if it’s tasteful. that wait time is usually when you’re still mentally in the task, not really in discovery mode, so it could feel distracting unless it’s extremely well targeted

the revenue share part helps though, makes it feel less like ads and more like a trade

also the privacy positioning is important, but people will still be skeptical by default, especially in dev tools

overall it clicks conceptually, just feels like execution and trust will make or break it more than the idea itself

The marketplace only works if both supply and demand scale together, developers won't stay opted in if the cards are irrelevant, and advertisers won't pay if developer engagement is low. Which side of that are you seeding first, and what's the honest state of the advertiser pipeline right now, are there actual paying advertisers or is the inventory currently being shown as house ads while you build the demand side?

This is a sharp observation about turning idle time into value—that's the same leverage play as turning signal quality into a scarce asset. Most multi-sided markets fail because they chase both sides simultaneously; the ones that win usually nail one side (developers here, it sounds like) before expanding. How are you thinking about supply/demand balance as you scale the advertiser side?

$1.58 per thousand quality attention seconds - so near minimum wage to watch advertisements?

nice to have you here

I like the idea because it monetizes dead time instead of creating another interruption. My biggest concern would be relevance. If I see low-quality suggestions a few times, I'd probably start ignoring the card completely. How are you deciding what each developer sees?

Congrats on the traction, Rizaldy. The idea clicks for me, but the thing that will decide adoption is the privacy claim, not the revenue split. You are asking developers to trust that something watching their AI workflow never touches code or prompts. That is a hard claim to just take on faith, most developers will want to see what signals it actually reads before installing.

I do product testing and walkthrough videos for early tools like this. Since you are already paying attention to this (revenue, real user feedback), I would frame it as a paid test cycle rather than a free pass: install BackSpin in a real dev workflow, document exactly what it observes versus what it claims to observe, and turn that into a video that doubles as trust-building content for future installs. Happy to talk specifics if useful.

I would love to write about your product. Can you let's discuss more about it?

I like that you're monetizing a moment that's usually ignored instead of competing for more attention.

The harder challenge will be preserving the feeling that developers are choosing to engage rather than being monetized. If that balance holds, the marketplace becomes much more valuable for both sides.