Tech and Startups 25-11-2025
Alarm bells ring over AI valuations; Databricks eyes $130B valuation; Revolut hits $75bn.
📣 Headlines
• Several investors and leaders are sounding the alarm on frothy AI valuations, warning the market may be overheating and posing systemic risks to tech companies and investors: investor calls AI 'dangerously overheated', DeepMind's Hassabis warns of a bubble and Google CEO Pichai cautions no company is immune.
• Some analysts push back that AI-related debt and hyperscaler buildouts are not a classic bubble, arguing bond markets reflect infrastructure growth rather than pure speculation (analysis).
• Mega-rounds and infrastructure bets continue: Databricks is reportedly in talks to raise at a $130B+ valuation, underscoring the broader theme that AI's next phase belongs to infrastructure.
• Large private-market activity persists with consumer-fintech strength as Revolut completes a share sale valuing it at $75bn, signalling continued appetite for late-stage rounds.
• Product teams should watch search behavior shifts—[consumers are moving from Google to ChatGPT], and startups like Peec AI raised $21M to help brands adapt to AI-powered search and GEO insights.
• Startups are productizing AI agents in verticals: security firms are shipping agent-native tooling with Secure.com debuting Digital Security Teammates while digital-health players scale AI therapy via M&A as Spectrum.Life buys Dr Jay to integrate an AI therapist.
• Talent and leadership are in flux as VCs prioritize hiring seasoned operators and execs to lead AI companies (VCs buying A-player C-suite talent), while top researchers exit big tech—Yann LeCun leaves Meta to start an independent AI entity and related coverage (BBC).
• Transportation and autonomy face a tougher funding environment: Crunchbase reports transportation dealmaking decelerating in 2025, and hardware-heavy startups show strain—Monarch Tractor warns of layoffs and a potential shutdown.
🔧 Company Engineering Blogs
Background Coding Agents: Context Engineering (Part 2) (engineering.atspotify.com). Context engineering for autonomous coding agents using Claude Code, prompts, and limited tools to migrate large codebases at Spotify
Evolving GitHub Copilot’s next edit suggestions through custom model training (github.blog). NES next-edit suggestions in GitHub Copilot trained with custom data, RL, and in-editor UX for VS Code
Generative UI: A rich, custom, visual interactive user experience for any prompt (research.google). Generative UI enables AI models to dynamically create immersive, customizable interfaces and tools for any prompt using Gemini 3 Pro in AI Mode
Making fairness in LLMs observable, quantifiable, and governable (amazon.science). FiSCo: a three-stage fairness evaluation for LLMs using controlled generation, semantic comparison, and statistical validation
👥 Team & Leadership
Ego, empathy, and humility at work (matthogg.fyi). Practical approach to ego, empathy, and humility for developers and technical leaders, emphasizing information pursuit and collaboration
Coding at work (after a decade away). (lethain.com). CTO Will Larson blogs on coding at work after a decade, AI tooling, small commits, and balancing management with hands-on software
The Making of a Decision - Part 2 (roundcrisis.com). Explores knowledge sharing, DIKW model, knowledge workers, and power dynamics in software teams
Adding People to a Late Project Will Make It Later (jjude.com). Explores Brooks’ Mythical Man-Month idea, arguing adding developers often delays projects; emphasizes architectural ownership and coherent system design
Invert, always invert (theengineeringmanager.substack.com). Invert pessimistically to plan, estimate, and roll out software projects with Charlie Munger’s inversion model
LGTM 🚀 Culture: A Short Story (alt.management). A speculative story criticizing LGTM culture, AI agents, and enterprise software hype through a personal, tech-history narrative
EOR vs Traditional Hiring Models: Strategic Guide for Global Tech Expansion (beontech.wpengine.com). Compare EOR and direct hiring for global tech expansion, focusing on speed, cost, and compliance across markets
Technical Advisor Playbook for 2026: What’s Changing and How to Adapt (diamantinoalmeida.com). Growth playbook for 2026: balance AI, automation, culture, and customer empathy with insights from a technical advisor
🚀 Startups & Growth
How Startups Get Viral (molodtsov.me). Viral startup success hinges on users, fanatic fan bases, exceptional product taste, and being unreasonably focused on a cornerstones idea
Context Engineering is the Secret: How Kuse Hit $10M ARR in 60 Days Without VC Funding (aakashg.com). Context engineering, visual context, Threads growth, Kuse 2.0, Xiankun Wu, Aakash interview, RAG, MVO vs MVP, 10M ARR in 60 days
🎧🍌 Inside the New Superhuman: $700M ARR, 40M Daily Users with Rahul Vohra (thespl.it). Rahul Vohra discusses Superhuman’s acquisition by Grammarly, Grammarly’s 40M DAU and $700M ARR, and building an AI-native productivity suite
Will VCs Lose Their Jobs to AI? (thisisgoingtobebig.com). AI reduces noise in sourcing; humans chase rare outliers in venture decisions with mentors like Lauren Reeves and Collin Wallace
So... I applied to Iterative. (geek.sg). Singapore-based founder shares applying to Iterative accelerator for Synscribe, covering AI-driven SEO, programmatic content, and revenue attribution insights
The Basecamp Blueprint: How to Bootstrap a Billion-Dollar Business (wearefounders.uk). Bootstrap a profitable tech empire by solving your own problems; Basecamp story of profitability, autonomy, and founder-led growth
I’m Open to Raising 200k for SaveOnTrading (kylevallans.com). Raising 200k–300k for SaveOnTrading to hire a full stack developer and scale revenue with media, trading, and tech expertise
📈 Product & Discovery
Solving Problems Worth Solving (candost.blog). Solving problems across software, psychology, and economics; leadership, design thinking, and managing complexity in teams
Cultivating the human capabilities that matter most (uxdesign.cc). Developing discovery judgment through psychological safety, cross-functional collaboration, and systematic practices for AI-enabled product discovery
Master Cross-Functional Collaboration Skills: The PM’s Guide to Career Acceleration (aakashg.com). Cross-functional collaboration for Product Managers: leading engineering, design, and GTM with frameworks, narratives, and an operating system
What are you optimizing for? (geek.sg). A startup-focused take on balancing value delivery and technical debt, emphasizing growth over perfection and stability in early stages
How to Avoid Solo Product Leadership Failure with a Product Value Team, Part 1 (jrothman.com). Explores product value teams as a remedy for solo product leadership, emphasizing collaboration, strategy, and experimentation
End-To-End with Shape Up: A Real-World Case Study (ryansinger.co). End-to-end Shape Up case study using framing, shaping, and build with a fractional CPO, senior engineer, SME, and a Notion kanban
How to work with product: To what port do you sail? (blog.nilenso.com). Explores aligning product and engineering toward clear goals, with examples from Simple.org and Gojek, and practical leadership tips
🤖 AI & Software Eng
Only you can stop AI database drops (stackoverflow.blog). How AI transforms software roles, guardrails for AI coding, and the rise of internal tools with Retool and Stack Overflow
How AI will change software engineering – with Martin Fowler (newsletter.pragmaticengineer.com). Martin Fowler discusses AI’s impact on software architecture, refactoring, non-determinism, vibe coding, and how tools like LLMs intersect with deterministic practices
I Never Really Loved Coding (And Only AI Made Me Realize It) (thebootstrappedfounder.com). AI-assisted coding shifts focus from writing to orchestrating code and business outcomes
Vibe Coding Myths vs. Reality: Why AI Isn’t a Bubble. Insights from llm-eval-simple (grigio.org). AI isn’t a tech bubble: local and frontier LLMs, prompt engineering, Rust vs Python, and Network Monitor demo
The Next Million Rails Apps (kodykendall.com). Rails boosts non-engineers with LLMs for browser-based AI coding, leveraging Ruby on Rails and Leonardo for maintainable, scalable apps
Hyperproductivity: The Next Stage of AI? (secondthoughts.ai). AI-powered teams build bespoke productivity tools, using Claude Code and Amplifier to automate product lifecycles and push recursive tool improvement
How I Built a Company AI Assistant in 4 Days (Without Being an AI Expert) (mindthenerd.com). Engineer builds on-premise company AI assistant in 60 hours using Ollama, vLLM, Docker, Ubuntu, C#, React; multilingual ERP integration
Good Enough and Really Fast: A Haiku 4.5 Case Study (barazany.dev). Fast, affordable AI coding with Haiku 4.5; Sensos teams cut costs and maintain flow using Cline and Claude Sonnet
📊 Product Metrics & Platforms
When To Kill A Project (blog.jacobstechtavern.com). Lessons from Dave Verwer on killing failing projects, market research, timing, and building open-source tools in iOS with Swift and Swift Package Index
The metrics product we built worked — But we killed it and started over anyway (blog.sentry.io). Sentry kills its first metrics product, shifts to trace-connected, event-based metrics with EAP and raw data storage for better debugging context
A RedMonk Conversation: David Mytton on Arcjet’s Vision for Developer Security (redmonk.com). Arcjet’s David Mytton discusses developer-focused security as a feature, WASM-based local-first security, and integrating bot detection, rate limiting, and email validation for modern JS/TypeScript stacks
[Link] An Interview with Cal Henderson (buildingslack.com). Cal Henderson discusses Slack's origin, Glitch shutdown, Flickr roots, engineering leadership, culture, and product philosophy with founder insights
Back to simplicity (eddiedale.com). Reflective pivot from Linux tinkering to simplicity, embracing focused web apps, 'good enough' products, and noteable creators
Linear vs. Jira (2025): Is Linear Finally Ready for Non-Technical Teams? (wearefounders.uk). Linear’s Form Templates and automation push position it as a potential Jira alternative for non-technical teams in 2025
🔭 Tech Trends & Essays
A startup in Mongolia translated my book (blog.pragmaticengineer.com). Mongolian startup Nasha Tech translates The Software Engineer's Guidebook, explores TokTok, Copilot, Claude Code, and Mongolian tech ecosystem
Fragments Nov 19 (martinfowler.com). Thoughtworks Technology Radar 33 focuses on AI wave, agents, infrastructure orchestration, coding workflows, and antipatterns with insights from Gergely Orosz and Siemens experts
Build to Last (oreilly.com). Software craftsmanship and AI; LLVM, Clang, Swift; insights from Chris Lattner and fast.ai founder Jeremy Howard
AI Emergence at ViteConf 2025: UX, DX, now AX (babbel.com). How AX emerges from UX and DX at ViteConf 2025, featuring Netlify’s CLI tweaks for AI agents, with Saheed Oladele and Evan You
“Productivity”. You Keep Using That Word. (notes.jim-nielsen.com). Quips on productivity, AI, web platform design, and career ethics from Jim Nielsen’s notes featuring Deming, Petros, Rupert, Carr, Russell, and Gaw
👋 Before you go...
I've got a big favor to ask - keeping Blaze running isn't expensive, but it does all add up, so I'm asking readers like you to help, if you can, by joining the Patreon page. Nothing flashy, just a way for folks who find value in these newsletters to chip in a little each month.
If you are getting value from blaze, checking this out would mean the absolute world. But if you can't contribute, no worries - the newsletters keep coming either way. Thanks for reading and being part of this nerdy corner of the internet. All the best for the coming week - Alastair.
Add a comment: