Executor v2 ships, DeepFinance launches long-horizon finance benchmark, and indie founders share distribution lessons

Key Takeaways

1. Executor v2 is positioned as a major rebuild aimed at code-driven Zapier/Retool-style workflows with full capability exposure to agents.
2. DeepFinance debuts as a long-horizon benchmark of 21 real-trading-data tasks built by finance and academic experts, each estimated at 20–40 human hours.
3. A builder's Notion-to-shelf project shows a low-cost, self-hosted approach to rendering a decade of reading data.
4. LeadOcean's AppSumo launch swung from $3,819 on day one to $84 on day two and back to $3,371 by day ten, with the founder citing distribution over polish.
5. An 18-year-old's AbMaxx shifted from UGC and influencer spend to Meta and TikTok ads, lifting monthly revenue from $10K to $60K.
6. levelsio shut down a 13-year-old Linode VPS serving 100+ sites and plans to publish a list of 10,000+ blocked AI reply accounts.
7. acquire.com listings include a profitable Lovable/Replit competitor at $304.2K TTM revenue and a $440K asking price.

1. Product Launches and Benchmarks

1. Executor v2 has been released and is described as a major product rebuild intended to cover all the use cases users had previously expected; its positioning is to make Zapier- and Retool-class capabilities fully code-driven and to expose all capabilities to agents 1.
2. DeepFinance launched as what its publisher calls the first ultra-long-horizon benchmark for high-end finance work, with 21 questions built on real trading data by experts with Evercore, JPM, UBS, HBS, and Wharton backgrounds; each question is estimated to take a human professional 20–40 hours. The publisher argues existing finance benchmarks only test document Q&A or simple numerical correctness and cannot cover end-to-end workflows full of assumptions and conflicting evidence 1.

2. Indie Founders on Distribution and Spend

1. Independent founder Ashish (LeadOcean) recorded $3,819 in AppSumo sales on launch day, $84 on day two, and $3,371 by day ten; Noah Kagan said day two is when most founders conclude a product has failed, while Ashish kept publishing content. Ashish's reflection was that he spent too long polishing the product instead of getting it to market early, and he kept posting on Reddit, private founder communities, and LinkedIn, saying "day two is the day to start posting, not stop" 1.
2. Ashish adjusted the product to actual usage: of 1,751 users, 1,358 never open the website and instead use LeadOcean inside Claude or ChatGPT, making it critical to define exactly what problem is being solved. He had quit a Shopify engineering job for dropshipping, lost everything during the pandemic, and restarted with no investment and no co-founder; he now runs LeadOcean largely alone at about 30 hours a week, with sales just past $150K, above his old salary. Kagan's takeaway: building products got cheaper, distribution did not 1.
3. Eighteen-year-old Lino built AbMaxx on Rork with no coding experience; after finding UGC and big-influencer partnerships squeezed margins to 5–10%, he cut them entirely and moved to Meta ads at 2–2.5x ROAS, keeping one equity-partner creator for content. He then moved the same creative to TikTok, where ROAS rose to 4–5x on stronger conversion among a younger audience, and the focused shift took monthly revenue from $10K to $60K 1.

3. Infrastructure, Platform Hygiene, and Small Builds

1. levelsio shut down "Tommy," a Linode VPS he had run for 13 years at $300/month hosting more than 100 sites; he said it served 300 million visits per year and 115,000 paying customers, and that he moved to Hetzner and split sites into isolated setups mainly for AI-era cybersecurity rather than performance, citing Linode's maintenance state after the Akamai acquisition. A collaborator said working with Tommy for nearly 12 years taught him Linux, systems administration, and performance optimization 1 2.
2. levelsio said he will publish a list of more than 10,000 blocked AI spam-reply accounts and submit it to X and third-party lists so others can bulk-block; he argued AI replies crowd out genuine interaction and that the growth tactic itself does not work. He said his reply sections quieted noticeably after he posted a warning and showed a self-built tool for blocking X reply bots 1 2 3 4 5.
3. A user built a personal reading shelf modeled on Ben's idea: ten years of reading records stored in Notion, pulled via API, with covers and page counts fetched free from Open Library without an API key. Spine width maps to page count (90 pages → 26px, 700 pages → 72px), color comes from the cover, and height is jittered by slug hash; covers are downloaded and self-hosted to satisfy CSP and avoid Open Library slowdowns. The builder had previously spent about $5 generating realistic spines for 140 books (roughly $0.03 each) but hit incorrect image proportions and needed to fix aspect ratio and prompts 1.

4. Market Listings and Deal Flow

1. acquire.com listed a SaaS startup positioned as a competitor to Lovable, Replit, Base44, and v0; it is profitable with $304.2K TTM revenue and a $440K asking price. Other listings include a fully automated AI SaaS for entertainment-industry casting submissions with $112K ARR run rate, $74.1K TTM revenue, and a $235K asking price; an AI fashion design platform with over 22,000 registered accounts, over 51,000 Instagram followers, $28.3K TTM revenue, and a $50K asking price; and a digital AI medical CME and medical social platform with $450K TTM revenue, 100% growth, and a $250K asking price 1 2 3 4.
2. Flippa shared Jeremy Schoemaker's advice that he kept each side project operating independently before selling, warning that shared PayPal accounts and a shared Google Workspace quickly lose buyer interest 1. Separately, acquire.com relayed the view that experienced founders should consider buying a second startup rather than building from scratch, for example acquiring a SaaS with $300K annual revenue to skip re-acquiring the first ten customers, while noting acquisition risks around customers, finances, and why the seller is selling 2.

5. Disputed/Unverified

1. levelsio relayed and agreed with a speculation that AI labs may be systematically reverse-engineering software, apps, and games on the internet with unguarded unreleased models, using reconstructed source code for training while building guardrails for themselves; the claim is presented as a guess without supporting evidence 1.
2. levelsio said Anthropic launched a personal-finance test feature for US users that connects accounts with read-only access and cannot move or invest funds, and argued AI labs will enter nearly every industry one by one; he cited building a spending tracker with Claude Code as support for that judgment 1 2.
3. levelsio said AI makes using a computer and staying indoors boring, and that many friends believe people should spend more time outdoors; he believes AI will help push people outdoors 1 2 3.