# Top 10 in Tech - What to know for Week ending October 17, 2025

Published: 16 October 2025
Canonical: https://www.top10in.tech/posts/week-ending-october-17-2025

## 1. SaaS METRIC OF THE WEEK: Forecasting

Forecasting MRR is more art than science - but [this post gives you both](https://www.thevccorner.com/p/saas-mrr-projection-model-excel-template). A smart walk-through of how to project MRR using cohort-based logic, with a FREEEEE Excel template that looks like it might actually work well.

Link: https://www.thevccorner.com/p/saas-mrr-projection-model-excel-template

## 2. GTM

Tomasz Tunguz just [dropped the GTM guide](https://tomtunguz.com/go-to-market-guide) every startup needs. Learn when to hire your first AE (after 10 founder-led deals is a great tech-ism), how to pick your sales model (field vs inside vs self-serve or a hybrid), and why the best GTM teams layer focus before they layer headcount. Includes pipeline targets, comp models, and team structures from $1M to $100M ARR.

Link: https://tomtunguz.com/go-to-market-guide

## 3. AI PRODUCT

There is a bit of OpenAI in this week's newsletter - so let's start here with an AI-ism. AI features aren't AI products. [This teardown of OpenAI's strategy](https://www.thevccorner.com/p/ai-product-strategy-openai-guide) shows what it takes to go beyond a wrapper: latency-aware UX, agent loops, invisible value, and goal-driven design. But shit moves so fast - that even [AI apps don't move as fast as the market](https://hex.tech/blog/bitter-lessons-building-ai-in-hex-product-management/).

Link: https://www.thevccorner.com/p/ai-product-strategy-openai-guide

## 4. OPENAI

OpenAI [Dev Day 2025 happened last week](https://www.scrumlaunch.com/blog/openai-devday-2025), and a few bombs dropped: GPT-5.5, [a native app SDK, and "Agents as a Service."](https://www.humai.blog/openai-dev-day-2025-the-one-more-thing-that-devoured-the-internet/) But the bigger shift? [Distribution](https://openai.com/index/introducing-apps-in-chatgpt/). A new AMD chip deal, first-party hosting, and native monetization rails signal OpenAI's ambition to own the full stack - infrastructure to app store and [operating system](https://techcrunch.com/2025/10/08/openais-nick-turley-on-transforming-chatgpt-into-an-operating-system).

Link: https://www.scrumlaunch.com/blog/openai-devday-2025

## 5. TALENT

Is there a [tech talent shortage](https://www.saastr.com/want-a-job-in-tech-sf-added-11k-net-new-hires-2x-ny-40x-austin-30x-miami/), or is it just a mismatch? The [WSJ reports that while layoffs have flooded the market](https://www.wsj.com/lifestyle/careers/in-a-sea-of-tech-talent-companies-cant-find-the-workers-they-want-76b7983a), companies still can't find the right skills for AI, security, and infrastructure.

Link: https://www.saastr.com/want-a-job-in-tech-sf-added-11k-net-new-hires-2x-ny-40x-austin-30x-miami/

## 6. MARKET

Check out the [GTMfund's take on the State of the Market:](https://thegtmnewsletter.substack.com/p/gtmfund-on-the-state-of-the-market) 75% of seed-stage startups have \<12 months runway, 60% of Series A+ aren't hitting growth targets, and 25% of sales teams are shrinking. Shheeeeeeeeet. Buyers are more risk-averse, and PLG is on the rise again - not for price, but for speed.

Link: https://thegtmnewsletter.substack.com/p/gtmfund-on-the-state-of-the-market

## 7. SALES 1

Early-stage selling is heavy on Founder-Led sales, but founders often get sales wrong because they confuse product belief with buyer readiness. [This playbook is great and covers how to build a sales process from scratch](https://foundersedition.co/p/the-early-sales-playbook), including market mapping, email scripts, handling objections, founder-led tactics, and why closing isn't the hard part - it's getting to the first meeting.

Link: https://foundersedition.co/p/the-early-sales-playbook

## 8. SALES 2

What do you do when growth stalls? [Jason Lemkin lays it out for us](https://www.saastr.com/when-growth-slows): double down on sales efficiency, revisit churn (see 10 below), and stop assuming marketing will save you. TL;DR: The answer is almost always "more pipeline."

Link: https://www.saastr.com/when-growth-slows

## 9. MULTI-LLM

I often run multiple LLM agents on one task just to verify output - I have a lot of hallucination trust issues emerging. [Simon Willison explores the rise of using agents in parallel for coding](https://simonwillison.net/2025/Oct/5/parallel-coding-agents) - as it's not just faster, but safer. Think multi-threaded prompts, error-checking via ensemble output, and LLM diversity as a helpful hedge against those friggin hallucinations.

Link: https://simonwillison.net/2025/Oct/5/parallel-coding-agents

## 10. CASE STUDY

A two-year mission to reduce churn. [This post breaks down what actually worked](https://www.toption.org/post/reflections-on-two-years-solving-churn) with onboarding tweaks, activation metrics, internal alignment (and what didn't). The big takeaway is that there are no silver bullets, just sustained efforts to figure it out**.**

Link: https://www.toption.org/post/reflections-on-two-years-solving-churn

## POD OF THE WEEK

Finishing it with a little more OpenAI, Reforge founder Brian Balfour on [why ChatGPT is on track to become the next major distribution platform](https://www.lennysnewsletter.com/p/why-chatgpt-will-be-the-next-big-growth-channel-brian-balfour).

Link: https://www.lennysnewsletter.com/p/why-chatgpt-will-be-the-next-big-growth-channel-brian-balfour