# Top 10 in Tech - What to know for Week ending August 28, 2026

Published: 28 August 2026
Canonical: https://www.top10in.tech/posts/top-10-in-tech-what-to-know-for-week-ending-august-28-2026

## 1. SaaS METRIC OF THE WEEK - MMM

This is a new acronym for your tech dictionaries - MMM, or [Marketing mix modeling](https://funnel.io/blog/marketing-mix-modeling-explained), is used to [measure the cross-channel impact](https://www.polestaranalytics.com/blog/marketing-mix-modeling-explained) of all the various marketing and non-marketing drivers on your startup's sales or business efforts and performance. Google's Meridian is its [open-source take on it](https://business.google.com/aunz/think/measurement/google-meridian-marketing-mix-modelling).

Link: https://funnel.io/blog/marketing-mix-modeling-explained

## 2. PRINCIPLES

I love me some [laws and principles](https://www.top10in.tech/posts/week-ending-august-1-2025) - here is a [new list of 10 of the lesser known ones](https://giacomofalcone.substack.com/p/the-principles-and-laws-youve-never). Hanlon's Razor [is a goodie](https://thedecisionlab.com/reference-guide/philosophy/hanlons-razor): rather than questioning people's intentions, question their competence.

Link: https://www.top10in.tech/posts/week-ending-august-1-2025

## 3. AI THESIS

Sarah Guo built Conviction around one bet: [the biggest AI opportunities sit above the foundation models](https://www.the-ai-corner.com/p/sarah-guo-conviction-ai-thesis), not inside them. Her logic is simple - the labs become infrastructure, while startups win by owning the application, workflow, and customer.

Link: https://www.the-ai-corner.com/p/sarah-guo-conviction-ai-thesis

## 4. VERTICAL INTEGRATION

Counterpoint to Sarah Guo above (#3). [Akash Bajwa reckons model labs and agent companies are converging](https://www.akashbajwa.co/p/the-endgame-of-vertical-integration) the stack, not separating - Anthropic is building products while Harvey is training models. The endgame may be owning the model, the application, the workflow, and the customer around it.

Link: https://www.akashbajwa.co/p/the-endgame-of-vertical-integration

## 5. WATERMARKS

Last week Anthropic/Claude [announced that it will now embed an invisible watermark into the text it generates](https://www.anthropic.com/news/claude-text-watermark) - and it can survive copy-paste and some editing. The weird bit: even using Claude to proofread or translate your own writing can leave the mark. Heavy rewriting can break it, but a third-party detection tool is coming. (not necessarily by choice - this is part of Anthropic's [EU AI Act compliance](https://businesschief.com/news/anthropic-watermarks-claude-to-meet-new-eu-ai-act-rules)).

Link: https://www.anthropic.com/news/claude-text-watermark

## 6. BENCHMARKS

Stop asking if you're above average. [This article](https://newsletter.getdx.com/p/your-benchmarks-dont-apply-to-us) argues the more useful benchmark question is whether you're improving faster than everyone or anyone else. Your own trend shows change; the market trend tells you whether that change was actually yours or just the tide coming in for everyone.

Link: https://newsletter.getdx.com/p/your-benchmarks-dont-apply-to-us

## 7. SPEND

Speaking of Benchmarks, the [SaaS Capital Spending Benchmarks for B2B SaaS Companies](https://www.saas-capital.com/wp-content/uploads/2026/06/Spending-Benchmarks-for-Private-B2B-SaaS-Companies-2026-W1.pdf) is out - bookmark this when it's time to benchmark your wallets y'all! Bootstrapped B2B SaaS companies spend 96% of ARR (hey - I do better than that!!), while VC-backed ones burn 101% (last year it was 95% vs 106%). Equity-backed teams in turn grow at 25% average vs 20% if your boots are strapped.

Link: https://www.saas-capital.com/wp-content/uploads/2026/06/Spending-Benchmarks-for-Private-B2B-SaaS-Companies-2026-W1.pdf

## 8. RETENTION

Meta laid off 10% of staff and reassigned 20-30% of software engineers to AI whatever - now they are reportedly [offering $400K-$1M+ in equity to senior engineers trying to resign](https://newsletter.pragmaticengineer.com/p/the-pulse-metas-self-inflicted-resignation-154) after layoffs/AI reassignments. Some still left for Google, Anthropic, and OpenAI.

Link: https://newsletter.pragmaticengineer.com/p/the-pulse-metas-self-inflicted-resignation-154

## 9. LIQUIDITY

Startup liquidity is quietly becoming its own market. [Carta administered 71 tender offers worth $3B in H1 2026](https://carta.com/data/tender-offer-update-h1-2026) - the highest first-half total it's logged in at least six years. With IPOs taking longer, tenders are becoming the liquidity pressure valve for employees and investors.

Link: https://carta.com/data/tender-offer-update-h1-2026

## 10. CASE STUDY

WISE: Wise's [original Seed deck](https://persuade-and-raise.beehiiv.com/p/wise-deck-breakdown-ce1c) is basically the opposite of most pitch decks today. No giant TAM slide, no AI-generated design theatre. It explains the problem, shows the before/after economics, proves people are already using it, then asks for $1.3M.

Link: https://persuade-and-raise.beehiiv.com/p/wise-deck-breakdown-ce1c

## POD OF THE WEEK

OpenAI's Head of Design Ian Silber on [why engineers have 10x'd with AI while design teams haven't](https://www.lennysnewsletter.com/p/openais-head-of-design-this-is-the), and why that makes design more important, not less.

Link: https://www.lennysnewsletter.com/p/openais-head-of-design-this-is-the