# "The first version didn't work, no excuses": Anchor founder turned an AI course into an endless puzzle feed

> Source: https://stringers.io/en/apps/play_com.oboe.android.html (Stringers / 什么值得做, generated 2026-11-08)
> Stringers answers one question every day: what small product can one or two people build in a few months and actually make money from. Numbers come from 15 databases we collect daily, each with source and date.
> For agents: these verdicts and the underlying databases (verified revenue, App Store / Google Play charts, ad libraries, GitHub, Product Hunt…) are queryable live over MCP — tools today / hot / products / directions / direction / data. Setup: https://github.com/komako-workshop/stringers (the user logs in at stringers.io → "配置 agent" to get a personal URL). Site index: https://stringers.io/llms.txt

> The founder posted a candid transition thread on X. Investor circles reshared it, and existing users plus TikTok accounts with 56k followers, warmed up in advance, sent traffic. The product itself hasn't spread on its own yet.

[Mobile App](https://stringers.io/en/apps/index.md) · Android · written 11-6

Oboe · Oboe Labs Inc

# "The first version didn't work, no excuses": Anchor founder turned an AI course into an endless puzzle feed

**What it is** An app that lets you scroll through brain-teaser games and mini trivia quizzes like short-form video. Each screen is generated by AI on the fly. Free, no ads.

**Why it got attention** The founder posted a candid pivot thread on X, which investors in his circle reshared. Thousands of existing users and a TikTok account with 56k followers, grown in advance, sent people over. The product itself hasn't spread on its own yet.

**C** **How you could follow** Don't copy the generic version head-on. Pick a smaller language or a narrow audience for a brain-teaser format that never ends. Copy its approach: use science explainers and daily game-list videos to build an audience, have comments drop keywords that trigger the link, and test paid tiers early.

Grade C: the early traction came from famous founders, the a16z circle and existing users, which followers can't copy. The same idea already has big players like Matiks and unseen indie versions, and it hasn't proven it makes money itself, so the playbook is what's worth learning. Buzzing in small circles

Data (updated daily) · first launch 2026-08-24 · 11-07 installs 9,555

Android · Total installs worldwide

Data 2026-10-29 → 2026-11-07 (10 days, 10 observations). Google's detail page only refreshes install counts every other day, so a flat day doesn't mean nobody installed.

Android · Ratings count (cumulative)

Data 2026-10-29 → 2026-11-07 (10 days, 10 observations)

In 2026 Nov 3, Nir Zicherman, who sold Anchor, posted a long thread on X admitting that Oboe, the AI course product he spent two years on, had failed. The same day he shipped a new version: an endless feed of brain-teaser mini games, pitched as a cure for doomscrolling and brainrot. A week later it still had fewer than 10k installs on Google Play. The numbers are small, but how it got here is worth a look: first figure out who is pushing it, then decide whether you can do it.

### A long thread admitting the mistake

Zicherman's post opens with: "We've pivoted Oboe. The first version… didn't work. No excuses." He then gave three reasons: no differentiation in the era of large models; users had no reason to come back because it was only useful; and the fatal problem: "No matter how good the product is, there just aren't many people who want to learn things every day on their own." The result was "a small group of loyal users who liked a product that never grew."

So they changed the motive: not relying on people wanting to learn, but on people wanting to be entertained. The new Oboe is a "fully personalized, dynamically generated, interactive anti-brainrot challenge feed", where each swipe is a puzzle or trivia question generated for you. The thread says beta users number "in the thousands" and that engagement and retention for the new version are "hugely different" from the old one. No third-party data can verify this.

The same day, co-founder Michael Mignano posted, and Matt Hartman (Factorial) reshared it: "The new @oboelabs is scrolling that makes you smarter".

### What it actually is

The store page has three lines: "A better way to pass the time. Finally, a real fix for doom-scrolling and brainrot. An endless interactive game and puzzle feed that feeds your brain." Free, no ads, no paywall. These are the three things it keeps stressing on TikTok.

Users agree on what they like. Five-star reviews from the US App Store 11-03: "I play it on the subway during my commute... it works smoothly even without signal" and "Between two calls on the subway, it feels better than endlessly scrolling Instagram or TikTok." A Play user 11-05 wrote: "I didn't think I'd like it. Please never add ads." The only real complaint came from 10-15: once you've scrolled through a game, you can't go back, and you have to find it in your favorites.

### Where the volume comes from: almost entirely the team's own resources

Google Play cumulative installs (public store number): 10-21 about 2,500,10-29 about 4,000,11-02 about 8,000,11-05 is 9,555. The fastest days were 10-31 to 11-02, with more than about 900 a day, right around the public launch. Afterward it fell back to about 400 a day. Ratings went from 10-29 at 60 to 11-06 at 128, adding 5 to 13 a day, very steady, with no sign of fake reviews. The iOS version "Oboe - Get smarter" only launched 10-31. As of 11-06, it had 19 ratings in the US store and was around #40 on the free Education chart.

On Play it reached #16 on the Canada Education free chart, #26 in Australia, #61 in the UK, and #87 in the US (11-06). That sounds good, but the English-speaking Education category charts have a very low bar, and a few hundred installs a day can get you on them.

Breaking it down, there are three sources, and followers can't get any of them:

| Source | What we saw |
|---|---|
| Founder and investor circles | Zicherman and Mignano posted it on X, Factorial's people reshared it, and the LinkedIn company page picked it up too. Old Oboe raised a 16M Series A led by a16z, with Garry Tan, Lenny Rachitsky and Adam D'Angelo following on (TechCrunch, 2025-12) |
| Existing users and beta testers | "Thousands of beta users" plus the users left over from the old product. Reviews on Play started in mid-October, and installs were already rising before the public launch. |
| A TikTok account grown in advance | @oboe_app 56k followers, detailed below |

No evidence of paid acquisition was found (the TikTok videos have no ad label, and Meta showed no ads from it). Media didn't cover it this time.

### The TikTok account is the part most worth copying

@oboe_app posted no product content before the pivot. 10-29 One science explainer about "crows going to find ants to treat themselves" got 195k views. The same day, another about an astronaut's space drift got 10k views. The other type was the "daily game list": 10-30 One post said "Comment GAMES to get these three plus links to hundreds of games" and got 7,700 views, tagged #wordle #connections #nytgames. 10-28 Another was "Connections editor and expert solving tips."

In other words, it first built an audience with general science content, then used "daily puzzle" content to pull in players of NYT Games, Wordle, and Connections. Comments with a keyword automatically DM the link. When the new product launched, this group was exactly the target audience.

The problem is that the product videos themselves didn't take off. 11-02 The first new-product video, "Introducing the all new Oboe", got 693 views, "puzzles that never end, no ads or paywall" got 455 views, and 11-05 that one got 276. The account's audience likes crows, not necessarily an app.

### Where the money comes from: nothing yet

It's completely free for now and makes no money, funded by investment (about $20M in total). The old web product had $15/month and $40/month subscription tiers (TechCrunch, 2025-12). The new version publicly promises no ads and no paywall, and says nothing about how it will charge later. Inference: either it adds a subscription once volume grows, or it uses retention data to raise the next round.

### Who it's competing with

"Replacing scrolling with brain teasers" isn't a new idea, and the category already has established players:

| Product | Scale | How it differs from Oboe |
|---|---|---|
| Matiks | Play 7.04M installs, +340k in a week; September 2.2M MAU; 10-26 just raised $10M (TechCrunch) | Math, logic, and memory puzzle battles, with ranked play and social features, also framed as "pulling people out of AI slop" |
| Puzzlo: Infinite Puzzles | 2026-09 launched on iOS, 5 ratings in the US store | Almost the same idea, built by an indie developer, and nobody noticed |
| Nibble | Play 2.96M installs | Daily trivia and quizzes, subscription |
| Elevate / Peak / Lumosity | About 29M Play installs each | A long-running brain-training app, a fixed game library plus subscription |
| Brilliant | Play 7.37M installs | A systematic STEM course, subscription |
| NYT Games / LinkedIn Games | Very large | A few questions a day, then it's over. Oboe is endless. |

Oboe's only new element is "every screen is generated by AI and adjusted to you." But this hasn't been proven as the reason it's growing. The current volume comes from the team's network and existing users. The Puzzlo example shows that one person can build "puzzles to replace scrolling" in two months. The hard part is getting people to see it.

### If it were you

**The short take:** Copy its playbook, but don't go head-to-head with this product. Oboe's traction comes from the Anchor founder's name, the a16z crowd and a few thousand early users, none of which you have. The same idea also has a big player in Matiks, and an unseen one in Puzzlo. And it still hasn't proven it makes money.

**Who it suits** : small teams that already run a content account or can make short videos, and have a specific type of brain-teaser in hand (word games in a particular language, geography or history quizzes for a country, or drill-style questions for a specific exam audience).

**Who it doesn't suit** : people who want to win on the tech point of "infinite AI-generated puzzles" but have no distribution channel.

**How to do it** : don't build a generic "anti-brainrot" app. Pick a language other than English or a narrow audience (for example, daily word games in Spanish or Japanese, or bite-sized exam drills for a specific test), and make it something that can be scrolled endlessly. Oboe promises it's free, so you can test subscriptions or ad-free paid tiers early to prove people will pay.

**How to acquire users** : copy its TikTok playbook. First build your target audience with general science content or "daily game list" posts, with comments dropping keywords that automatically send the link. When you launch, post a candid development-story thread. Minority-language Education charts have a low bar, and a few hundred installs can get you on them, which picks up some store traffic along the way.

**How much time you have** : there's no hard window. "Brainrot alternative" is a hot phrase right now, and Matiks just raised money. Inference: more similar products will appear within six months, so the sooner you claim a language or audience, the better.

**What would mean you read it wrong** : if Oboe grows on its own within a month or two without relying on resources (new-product videos start hitting six-figure views, and Play daily installs stabilize at over a thousand), that would show the "AI-personalized puzzle feed" has real spread power and is worth following more directly. Conversely, if it quickly adds a paywall and retention drops, that would show the format can't hold users, and even the playbook is worth less than it looks.

[Android store page ↗](https://play.google.com/store/apps/details?id=com.oboe.android)

**Research notes: competitor table, acquisition evidence, full research report**

Competitors

| Name | Where | Launched | Scale | How it differs from this one |
|---|---|---|---|---|
| Matiks | Android | 2024-08 | Play 7.04M installs, +340k in 7 days; 2.2M MAU; raised $10M 2026-10 | Math/logic puzzle battles with ranked and social features; Oboe is a single-player, AI-personalized, endless information feed |
| Puzzlo: Infinite Puzzles | iOS | 2026-09-03 | 5 ratings on the US iOS store | Almost the same idea (puzzles to replace scrolling), built by an indie developer with no resources and no AI personalization |
| Nibble | Android | 2022(iOS)/2025(Play) | Play 2.96M installs | A subscription-based daily trivia and quiz app; Oboe leans toward games and is free with no ads |
| Imprint | iOS | 2019 | iOS US store 60k ratings | Illustrated bite-sized learning, subscription |
| Elevate / Peak / Lumosity | Android | 2014-2015 | About 29M Play installs each | Fixed brain-training game library plus subscription, not an endless feed |
| Brilliant | Android | 2014 | Play 7.37M installs | Systematic STEM courses, subscription; Oboe dropped courses and pivoted to entertainment |
| NYT Games / LinkedIn Games | Big-company features | — | Very large | A fixed few questions a day, done in one sitting; Oboe is endless scrolling |

Where users come from

- **Verified** **by founders and investors posting on X/LinkedIn** : Zicherman 11-03 long transition post, reposted and posted the same day by Mignano and Matt Hartman (Factorial)

- **Verified** **Beta and migration of old users** : the pivot post mentions thousands of beta users; the old product had a small loyal user base; installs started climbing in late October, when the beta began

- **Verified** **Self-run TikTok account with comment-to-get-link** : @oboe_app with 55.9k followers. Before the pivot it posted science explainers and a daily games list (up to 195k views), using Comment GAMES/OBOE for auto-DM links. New product videos get 266-690 views

- Inferred **Micro-influencer UGC** : @sam.psycho.smith 11-05 recommended Oboe, 202 views

- Inferred **Paid acquisition** : no Oboe ads found in the metaads library (the library doesn't track it); TikTok videos have no ad label

- Inferred **Store category charts** : Education Top Free charts in Canada, Australia, the UK, and US 16/26/61/87, a low barrier to entry

## Oboe (com.oboe.android): why it's breaking out

**Bottom line** : this isn't a small app built from scratch. It's a **pivot** by Oboe Labs, a company a16z has invested in (founded by the two Anchor co-founders, with about $20M raised in total). The original AI-generated course product never took off, and 2026-11-03 publicly pivoted to "an endless brainrot-free puzzle feed." The current traffic (about 9,555 Play installs) comes from three places: the founders' and investors' circles posting on X and LinkedIn, the few thousand beta users from before the pivot, and the TikTok account @oboe_app that the team built themselves (56k followers). Volume is still small, and "anti-doomscrolling puzzles" already has big players like Matiks (7M installs, recently raised $10M).

### 1. Data (from our database)

- Google Play cumulative global installs (googleplay library app_day / app_day_ext): 10-21 is 2,533,10-26, 4,000,10-30 is 5,161,11-02, 8,055,11-05 is 9,555. Up 4,394 in the last 7 days. Per AppstoreSpy's backfill, daily adds were about 10-21 at 1,184, then 250–400 per day after that; from 10-31 to 11-02 it averaged about 965,11-03 per day, and from 686,11-04/05 about 407 per day.

- Rating count went from 10-29 at 60 to 11-06 at 128, +5–13 per day, a steady pace with no sign of coordinated fake reviews (126 reviews on the Play page).

- Education Top Free chart: Canada 16, Australia 26, UK 61, US 87 (data card 11-06). The education category in these English-speaking stores is an easy chart to crack; a few hundred installs in a day can get you on it.

- The Android listing date in the store is 2026-08-24 (possibly the old product's Android version, or the date was reset when the info was changed). But the backfill curve only starts in late October, so there was basically no volume before October.

- iOS is a new standalone listing: "Oboe - Get smarter" (id 6794063907), live since 2026-10-31. As of 11-06, the US store had only 14 ratings, ranking #40–43 on the Education Top Free chart; #60 on New Zealand's Education Top Free and #120 in Singapore (appstore library). The iTunes API shows 16 ratings with an average of 4.75.

- Monetization: the TikTok profile and video captions say "100% ad free" and "no paywall"; the Play data card has no IAP range. **It isn't making money right now** , and is funded by fundraising. The old web product had two subscription tiers at $15/month and $40/month (TechCrunch 2025-12-10).

### 2. Who they are and why now

- Oboe Labs was founded by Nir Zicherman and Michael Mignano, co-founders of Anchor, which was later acquired by Spotify. 2024-10 raised a $4M seed round led by Eniac, 2025-09 launched an AI-generated course platform, and 2025-12 raised a $16M Series A led by a16z 16. Individual participants included Garry Tan, Lenny Rachitsky, Adam D'Angelo, and others (TechCrunch 2025-12-10).

- **2026-11-03 publicly pivots** . Zicherman's long X post (x.com/NirZicherman/status/2117683979809890720), roughly: "We spent two years on the first version and it didn't work... In the LLM era we had no differentiation, users had no reason to come back, and not many people actively want to learn every day." The new direction is "a fully personalized feed of dynamically generated, interactive, anti-brainrot challenges." The post also says "thousands of beta users," and that engagement and retention for the new vs. old versions were "dramatically different."

- The timing lines up: installs started climbing in late October, during the beta and soft launch; around the public launch of 11-03 (10-31 to 11-02), daily adds spiked to about 965.

- The broader backdrop: "brainrot/doomscrolling" has been a trending term for 2025–26 years, and TikTok is full of "anti-brainrot apps" list videos. Matiks raised funding in 10-26 with a similar pitch of pulling people out of the AI slop timeline (TechCrunch 2026-10-26).

### 3. Where users come from (growth tactics)

| Channel | Evidence | Confirmed? |
|---|---|---|
| Founder and investor networks post on X/LinkedIn | Zicherman's long-form transition post 11-03; Mignano posted the same day (x.com/mignano/status/2117685298603737106); Factorial's Matt Hartman reposted it ("The new @oboelabs is scrolling that makes you smarter", x.com/MattHartman/status/2117765474368864361); the LinkedIn company page published it too | Confirmed |
| Old-user migration and beta | The post mentions "thousands of beta users," and the old product had a "small, loyal user base" | Confirmed (numbers not verified) |
| Self-run TikTok account @oboe_app | 55.9k followers, 223 videos, 929k likes in total. Pre-pivot content was science explainers (the crow video at 195k views, 10-29) and a "daily games list" (10-30 at 7,505 views). Captions say "Comment GAMES / Comment OBOE for the link," and comments trigger an automatic DM with the link (ManyChat-style), pulling in NYT Games / Connections / Wordle players. New post-pivot product videos get only 266–690 views | Confirmed (TikHub) |
| Micro-influencers and UGC | @sam.psycho.smith 11-05 posted "I forgot to mention Oboe!!" (202 views), a small lifestyle account, likely paid or swap UGC; very small reach | Partly confirmed |
| Paid acquisition | Oboe wasn't found in the Meta ad library by page, play_id, iOS id, or domain (metaads only tracks products on its list, so not finding it doesn't mean it isn't advertising); TikTok videos have no is_ads flag | No evidence found |
| Press | No TechCrunch or other media coverage found for this pivot (the old version got coverage from TechCrunch and 2025-09 2025-12) | None for now |
| Store search | Small-country education charts have rankings and will bring in a bit of organic traffic | guessing |

### 4. Competitors

| Name | Where | Launched | Scale | How Oboe differs |
|---|---|---|---|---|
| Matiks | iOS/Android | 2024-08(Play) | Play: 7.04M installs, +340k in 7 days; 300k DAU, 2.2M MAU (2026-09); just raised $10M (General Catalyst) | Math/logic/memory puzzle battles with ranked play and social features; Oboe is a single-player, personalized, AI-generated feed with no battles |
| Puzzlo: Infinite Puzzles | iOS/Android | 2026-09-03(iOS) | iOS US store 5 ratings, indie developer | Same idea, "scrolling visual puzzles to replace doomscrolling," nearly the same pitch, but without the resources |
| Nibble | iOS/Android | 2022–25 | Play: 2.96M installs; iOS 47k ratings | Daily trivia and quizzes, subscription model; Oboe leans toward games and puzzles, free with no ads |
| Imprint | iOS/Android | 2019 | iOS 60k ratings, #7 on the US Education Top Free chart | Illustrated bite-sized learning, subscription; content is reading-based, not games |
| Elevate / Peak / Lumosity | iOS/Android | 2014–15 | About 29M Play installs each | Long-standing brain-training app with a fixed game library plus subscription; Oboe's pitch is "endless feed + AI dynamic generation" |
| Brilliant | iOS/Android/web | 2014 | Play 7.37M | Systematic STEM courses, subscription; Oboe dropped "courses" and switched to entertainment |
| NYT Games / LinkedIn Games | Big-company features | — | Very large | A fixed few questions a day, done in one sitting; Oboe is endless scrolling |
| AI game-generation platforms like Google Playground | Big-company features | 2026-10 | Experimental stage | General-purpose game generation; Oboe focuses on brain teasers only |

### 5. Differentiation

- The pitch has three points: the endless-scroll interaction, where every screen is a personalized brain teaser or trivia question generated dynamically by AI; completely free, with no ads. It maps TikTok's swipe gesture onto the daily-puzzle habit of NYT Games players.

- This differentiation **isn't why it's breaking out** . Current installs are about 10k, mostly from the founding team's resources (network, beta users, a TikTok account set up in advance), not from the product spreading on its own. New product videos get only a few hundred views, so organic word of mouth hasn't shown up yet. Puzzlo also does "puzzles instead of doomscrolling," and an indie developer could build that in two months. The product-level moat is the quality and personalization of AI-generated content, which hasn't been proven yet.

### 6. How many people have seen it

- Chatter stays in small circles: only founders, investors, and VC people discussing it on X/LinkedIn. The pivot hasn't been covered by the media, and there's no public revenue (it's free for now). The old Oboe's funding is public news (TechCrunch).

- Copycats: the category is already crowded (Matiks, Puzzlo, a pile of anti-brainrot list apps), but no one has been seen directly copying Oboe's new version.

### 7. Takeaways

- Before launch, grow a TikTok account around general science and "daily games list" content, then funnel traffic to the new product at launch; use keyword-triggered auto-DMs to turn video viewers into downloads.

- Position directly on a hot keyword: "doomscrolling/brainrot alternative." The user logic is "swap guilty scrolling for scrolling that pays off."

- A candid pivot post ("the first version didn't work, no excuses") spreads easily on VC X, a low-cost way to cold-start.

### 8. Uncertain

- Whether the beta users are real and how many old users migrated over; whether the Android release date 08-24 is the date of the older app.

- Whether they run paid acquisition on Meta/TikTok (our ad libraries don't track it); their future monetization (they currently promise no ads, no paywall).

- Whether retention is really as good as the founder claims (no third-party data).

### Sources

- googleplay library app_day / app_day_ext (2026-10-21 to 11-06); appstore library app_profile / app_day (6794063907); metaads library ad table (query returned empty)

- [x.com/NirZicherman/status/2117683979809890720](https://x.com/NirZicherman/status/2117683979809890720)(2026-11-03)

- [x.com/mignano/status/2117685298603737106](https://x.com/mignano/status/2117685298603737106) ;[x.com/MattHartman/status/2117765474368864361](https://x.com/MattHartman/status/2117765474368864361)

- [techcrunch.com/2025/12/10/oboe-raises-16-million-from-a16z-f…](https://techcrunch.com/2025/12/10/oboe-raises-16-million-from-a16z-for-its-ai-powered-course-generation-platform/)

- [techcrunch.com/2025/09/10/after-selling-to-spotify-anchors-c…](https://techcrunch.com/2025/09/10/after-selling-to-spotify-anchors-co-founders-are-back-with-oboe-an-ai-powered-app-for-learning/)

- [techcrunch.com/2026/10/26/matiks-raises-10m-to-expand-its-ma…](https://techcrunch.com/2026/10/26/matiks-raises-10m-to-expand-its-math-and-puzzle-gaming-platform/)

- TikHub: @oboe_app profile and videos, @oboelabs, @sam.psycho.smith (2026-11-06 scraped)

- iTunes lookup 6794063907 / 6772423511(Puzzlo)

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.