# Tatenda Gatawa

> Building with AI and data. What I'm reading and who I follow, refreshed every morning,
> plus the contract work I take on. https://tatendagatawa.dev

Updated 26 October 2026.

## Reading (newest first)

- [In Pictures: Securing Your Business In The Age Of Shadow AI NZ event with The Instillery (VITG)](https://www.itnews.com.au/gallery/in-pictures-securing-your-business-in-the-age-of-shadow-ai-nz-event-with-the-instillery-vitg-629897): iTnews, 26 Oct
- [IAG's SAP migration brings process and data into sharp focus](https://www.itnews.com.au/news/iags-sap-migration-brings-process-and-data-into-sharp-focus-629812): iTnews, 26 Oct
- [New IDC research reveals the real AI readiness gap: It's not models, it's data](https://www.itnews.com.au/feature/new-idc-research-reveals-the-real-ai-readiness-gap-its-not-models-its-data-629532): iTnews, 26 Oct
- [Data Engineering Weekly #293](https://www.dataengineeringweekly.com/p/data-engineering-weekly-293): Data Engineering Weekly, 26 Oct
- [A Flood of Lawsuits Could Help Make AI Safer](https://www.luizasnewsletter.com/p/a-flood-of-lawsuits-could-help-make): Luiza's Newsletter, 26 Oct
- [The five components of building a "company harness"](https://www.langchain.com/blog/components-building-company-harness): LangChain Blog, 26 Oct
- [Australia swaps out its ambassador for cyber and AI as global tensions mount](https://www.themandarin.com.au/322905-australia-swaps-out-its-ambassador-for-cyber-and-ai-as-global-tensions-mount/): The Mandarin, 26 Oct
- [Scott’s Tots and the Broken Promise Economy](https://joereis.substack.com/p/scotts-tots-and-the-broken-promise): Joe Reis, 26 Oct
- [Full Tutorial: How to Make Amazing Videos with Claude and HyperFrames | Jake Moran](https://creatoreconomy.so/p/how-to-make-amazing-videos-with-claude-and-hyperframes-jake-moran): Behind the Craft, 26 Oct
- [Podcast: Kimi K3.1 Pro, the new cyber reality, and Muse musings](https://www.interconnects.ai/p/podcast-kimi-k31-pro-the-new-cyber): Interconnects, 25 Oct
- [Labor Replacement Is AI’s Small Prize](https://www.gettheleverage.com/p/labor-replacement-is-ais-small-prize): The Leverage, 25 Oct
- [Frank and Fearless 2026: Most public servants using AI at work](https://www.themandarin.com.au/322881-frank-and-fearless-2026-most-public-servants-using-ai-at-work/): The Mandarin, 25 Oct
- [🔮 The internet under attack, AGI compression & the next oil shock that never will be... #607](https://www.exponentialview.co/p/ev-607): Exponential View, 25 Oct
- [Building Reasoning Models from the Ground Up: A Full Video Course](https://magazine.sebastianraschka.com/p/building-reasoning-models): Ahead of AI (Sebastian Raschka), 24 Oct
- [The State of Context Engineering](https://www.dbreunig.com/2026/10/24/the-state-of-context-engineering.html): Drew Breunig, 25 Oct
- [[AINews] Kimi 3.1 Pro 104BA-2.8T: Fable-5.1-Class Open Weights narrows Closed-AI lead to 1 month... and Pro Highspeed API with 2-3× speed](https://latent.space/p/ainews-kimi-31/): Latent Space, 24 Oct
- [battledrones v14: Compass Rose, a four player map](https://simonwillison.net/2026/Oct/24/battledrones/): Simon Willison's Weblog, 24 Oct
- [Context Engineering with Deep Agents: How To Remember Without Keeping Everything In Context](https://www.langchain.com/blog/context-engineering-deep-agents-remember): LangChain Blog, 24 Oct
- [Power Apps: Model-driven apps - Use row summaries in US Government clouds](https://www.microsoft.com/microsoft-365/roadmap?id=576279): Microsoft 365 Roadmap, 24 Oct
- [Dynamics 365 Sales: Lead management - Personalize outreach emails generated by Sales Qualification Agent](https://www.microsoft.com/microsoft-365/roadmap?id=578077): Microsoft 365 Roadmap, 24 Oct
- [Dynamics 365 Sales: Lead management - Deploy multiple Sales Qualification Agents in a single environment](https://www.microsoft.com/microsoft-365/roadmap?id=578076): Microsoft 365 Roadmap, 24 Oct
- [Quick Study](https://simonwillison.net/2026/Oct/23/quick-study-painter/): Simon Willison's Weblog, 24 Oct
- [Context Engineering (Agent Memory) for AI Agents Starts with Retrieval](https://www.langchain.com/blog/context-engineering-starts-with-retrieval): LangChain Blog, 24 Oct
- [How Finance Cut Forecasting to Hours With AI](https://www.snowflake.com/content/snowflake-site/global/en/blog/sales-finance-ai-forecast): Snowflake Blog, 24 Oct
- [Far from RSI? Compute Shortages? Google CEO Sundar Pichai’s View Of AI’s Trajectory](https://www.bigtechnology.com/p/far-from-rsi-compute-shortages-google): Big Technology, 24 Oct
- [Introducing Kimi K3.1 Pro](https://simonwillison.net/2026/Oct/23/kimi-k31-pro/): Simon Willison's Weblog, 24 Oct
- [The Epoch Brief - October 23, 2026](https://epochai.substack.com/p/the-epoch-brief-october-23-2026): Epoch AI, 24 Oct
- [Built for Agents: The Case for Spark Declarative Pipelines](https://www.databricks.com/blog/built-agents-case-spark-declarative-pipelines): Databricks Blog, 24 Oct
- [Real-time energy trading and analytics on Databricks](https://www.databricks.com/blog/real-time-energy-trading-and-analytics-databricks): Databricks Blog, 24 Oct
- [Bringing foundational models to the front lines of TB screening](https://research.google/blog/bringing-foundational-models-to-the-front-lines-of-tb-screening/): Google Research Blog, 24 Oct
- [The internet’s Wuhan moment](https://www.exponentialview.co/p/the-internets-wuhan-moment): Exponential View, 24 Oct
- [Multi-Turn Reinforcement Learning on Amazon SageMaker AI: A hands-on guide](https://aws.amazon.com/blogs/machine-learning/multi-turn-reinforcement-learning-on-amazon-sagemaker-ai-a-hands-on-guide/): AWS Machine Learning Blog, 24 Oct
- [Rapidly scaling online storage to serve over 1 billion ChatGPT users (Part II)](https://openai.com/index/scaling-storage-one-billion-users-part-two): OpenAI News, 24 Oct
- [Orchestrating SageMaker AI endpoint auto scaling with AWS CDK](https://aws.amazon.com/blogs/machine-learning/orchestrating-sagemaker-ai-endpoint-auto-scaling-with-aws-cdk/): AWS Machine Learning Blog, 24 Oct
- [How Bayer Crop Science built an agentic decision engine on Amazon Bedrock AgentCore with AWS Unified Operations](https://aws.amazon.com/blogs/machine-learning/how-bayer-crop-science-built-an-agentic-decision-engine-on-amazon-bedrock-agentcore-with-aws-unified-operations/): AWS Machine Learning Blog, 24 Oct
- [LitPivot: Developing research ideas through dynamic contextualization within the literature landscape](https://huggingface.co/blog/allenai/litpivot): Hugging Face Blog, 24 Oct
- [AI Is Getting Good at Math. Here’s Why.](https://latent.space/p/ai-is-getting-good-at-math/): Latent Space, 24 Oct
- [Clouded Judgement 10.23.26 - AI Goes Vulnerability Hunting](https://cloudedjudgement.substack.com/p/clouded-judgement-102226-ai-goes): Clouded Judgement, 23 Oct
- [Can data agents do real analysis?](https://roundup.getdbt.com/p/can-data-agents-do-real-analysis): The Analytics Engineering Roundup, 23 Oct
- [Take actions on your notebooks from Gemini Notebook directly in the Gemini app](http://workspaceupdates.googleblog.com/2026/10/take-actions-on-your-notebooks-from-Gemini-Notebook-directly-in-the-Gemini-app.html): Google Workspace Updates, 24 Oct
- [[AINews] Black October](https://latent.space/p/ainews-black-october/): Latent Space, 23 Oct
- [How to Measure AI Value Beyond Adoption Metrics | Snowflake](https://www.snowflake.com/content/snowflake-site/global/en/blog/measure-ai-value-beyond-adoption): Snowflake Blog, 24 Oct
- [Securing our frontier model training environment](https://openai.com/index/securing-our-frontier-model-training-environment): OpenAI News, 23 Oct
- [‘It takes a mess and amplifies it’: Five AI finance mistakes businesses should avoid](https://www.smartcompany.com.au/partner-content/five-ai-finance-mistakes-businesses-should-avoid/): SmartCompany: Artificial Intelligence (Neural Notes), 23 Oct
- [Canva’s first public image generation model cements AI ambitions](https://www.smartcompany.com.au/artificial-intelligence/canvas-first-public-image-generation-model-cements-ai-ambitions/): SmartCompany: Artificial Intelligence (Neural Notes), 23 Oct
- [How Large is the AI Circle?](https://tomtunguz.com/how-large-is-the-ai-circle/): Tomasz Tunguz, 23 Oct
- [Serving DeepSeek-V4.1-Flash at Scale: vLLM Matches DeepInfer and Scales to GB300 NVL72](https://vllm.ai/blog/2026-10-23-deepseek-v41-flash-pd-wide-ep): vLLM Blog, 23 Oct
- [GLM-5.3 Optimizations, Part 2: PCP, DCP, and Hybrid HiSparse on 16 H200s](https://vllm.ai/blog/2026-10-23-glm53-part2-pcp-dcp-hisparse): vLLM Blog, 23 Oct
- [vLLM Delivers Day-0 Support for MiniMax-M3.1-Flash](https://vllm.ai/blog/2026-10-23-minimax-m31-flash): vLLM Blog, 23 Oct
- [Does Your Model Run on AWS Trainium? Look It Up](https://huggingface.co/blog/torch-neuronx-model-support): Hugging Face Blog, 23 Oct
- [OpenAI and Poland’s NASK partner to advance AI safety and security](https://openai.com/index/openai-nask-ai-safety-security-partnership): OpenAI News, 23 Oct
- [Turning Database Grading into RL Rewards: Post-Training Agents with ThinkingBox](https://huggingface.co/blog/microsoft/thinkingbox-training): Hugging Face Blog, 23 Oct
- [Expanding Claude availability and deployment choice on Google Cloud](https://cloud.google.com/blog/products/ai-machine-learning/expanding-claude-availability-on-google-cloud/): Google Cloud Blog: AI & Machine Learning, 23 Oct
- [Your October AI Forum update: Vote, connect and explore what’s next for AI in Aotearoa](https://aiforum.org.nz/2026/10/23/your-october-ai-forum-update-vote-connect-and-explore-whats-next-for-ai-in-aotearoa/): AI Forum New Zealand, 23 Oct
- [The new economics of AI cost management](https://www.databricks.com/blog/ai-cost-management): Databricks Blog, 23 Oct
- [How to build a software factory that tests its own prompts](https://arize.com/blog/how-to-build-a-software-factory-that-tests-its-own-prompts/): Arize AI Blog, 23 Oct
- [The Internet's AI Security Wake-Up Call](https://www.luizasnewsletter.com/p/the-internets-ai-security-wake-up): Luiza's Newsletter, 23 Oct
- [Google Vault now supports filtered exports for Gemini in Workspace](http://workspaceupdates.googleblog.com/2026/10/google-vault-now-supports-filtered-exports-for-Gemini-in-Workspace.html): Google Workspace Updates, 23 Oct
- [Public Sector Data Modernization for AI Readiness](https://www.snowflake.com/content/snowflake-site/global/en/blog/public-sector-data-modernization-ai-readiness): Snowflake Blog, 23 Oct
- [Beyond tree cover: mapping the world’s forest types](https://deepmind.google/blog/beyond-tree-cover-mapping-the-worlds-forest-types/): Google DeepMind Blog, 23 Oct

## AI adoption and enablement

Most AI programmes stall on people, not technology. Who to read on rolling AI out across a workforce, building the enablement function, and measuring outcomes rather than logins.

### Who to follow

- [One Useful Thing (Ethan Mollick)](https://www.oneusefulthing.org/): Evidence on how AI changes work and organisations, written for leaders
- [The Every Podcast](https://every.to/podcast): Weekly conversations with operators putting AI to work inside companies, including Every's enterprise adoption practice
- [Elena Verna](https://www.elenaverna.com): Writes plainly about how AI is changing jobs, teams and growth work
- [Lenny's Newsletter](https://www.lennysnewsletter.com): The largest product community; its AI surveys and practitioner interviews show what teams actually adopt
- [Engineering Enablement (DX)](https://newsletter.getdx.com): Evidence on AI productivity in engineering teams, measured rather than claimed
- [MIT Sloan Management Review](https://sloanreview.mit.edu/): Research-backed articles on operating models, leadership and getting value from AI
- [Behind the Craft](https://creatoreconomy.so): Hands-on tutorials showing non-engineers how to prototype and work with AI tools
- [Wharton Knowledge: AI](https://knowledge.wharton.upenn.edu/special-report/2025-ai-adoption-report/): Publishes the annual enterprise AI adoption report and research on AI at work

### Start with these

- [Navigating the Jagged Technological Frontier](https://www.hbs.edu/faculty/Pages/item.aspx?num=64700): Dell'Acqua, Mollick et al. (Harvard Business School working paper). Field experiment with BCG consultants showing where AI lifts performance and where it hurts
- [AI tools are overdelivering: results from a large-scale AI productivity survey](https://open.substack.com/pub/lenny/p/ai-tools-are-overdelivering-results): Lenny's Newsletter. What practitioners say AI actually changed in their work
- [2025 AI adoption report](https://knowledge.wharton.upenn.edu/special-report/2025-ai-adoption-report/): Wharton. How large US firms are using and measuring generative AI
- [AI adoption is a UX problem](https://thenanyu.com/ux.html): Nan Yu. Why rollouts stall on interface and habit, not capability
- [Guidance for AI Adoption](https://www.industry.gov.au/publications/guidance-for-ai-adoption): National AI Centre / Department of Industry, Science and Resources. Current government guidance on the practices organisations should put in place when adopting AI
- [Unlocking productivity with generative AI: evidence from experimental studies](https://www.oecd.org/en/blogs/2025/07/unlocking-productivity-with-generative-ai-evidence-from-experimental-studies.html): OECD. A sober summary of measured productivity gains

## Agentic workflow automation

Agents that do back-office work beside the systems you already run: claims, lending, finance operations, service. How they are built, where they break, and how to keep a human in the loop.

### Who to follow

- [Anthropic Engineering](https://www.anthropic.com/engineering): First-party write-ups on agent design, tool writing, context engineering and multi-agent systems from people who ship them
- [Model Context Protocol Blog](https://blog.modelcontextprotocol.io/): Spec changes, roadmap and security guidance for the protocol most agent tool integrations now use
- [LangChain Blog](https://www.langchain.com/blog): Regular posts on agent architectures, agent evaluation and production case studies
- [Microsoft Copilot Studio Blog](https://www.microsoft.com/en-us/microsoft-copilot/blog/copilot-studio/): What is new for building agents in Copilot Studio, with governance and licensing changes
- [Designing with AI (Victor Dibia)](https://newsletter.victordibia.com/): Focused on multi-agent design patterns and how to evaluate them
- [Drew Breunig](https://www.dbreunig.com/): Clear framing of how context, prompts and tools fit together in agent systems
- [Philipp Schmid](https://www.philschmid.de/): Short, code-heavy guides on agent frameworks, function calling and managed agents
- [Foundation Capital](https://foundationcapital.com/system-of-agents/): Coined 'service as software' and the system-of-agents thesis for enterprise AI

### Start with these

- [Building effective agents](https://www.anthropic.com/research/building-effective-agents): Anthropic. The standard reference for when to use workflows versus agents, with the core patterns named
- [Effective context engineering for AI agents](https://www.anthropic.com/research/effective-context-engineering-for-ai-agents): Anthropic. Explains how to manage what goes into an agent's context window over long tasks
- [Agentic AI in organisations: early insights from practitioner interviews](https://oecd.ai/en/ai-publications/agentic-ai-in-organisations): OECD. How organisations are deploying agents and what governance they need
- [How we built our multi-agent research system](https://www.anthropic.com/research/multi-agent-research-system): Anthropic. A production multi-agent architecture with honest notes on cost, prompting and evaluation
- [Don't Build Multi-Agents](https://cognition.com/blog/dont-build-multi-agents): Cognition (Walden Yan). The strongest counter-argument: why shared context beats splitting work across agents
- [Owning the workflow in B2B AI apps](https://a16z.com/owning-the-workflow-in-b2b-ai-apps/): a16z. Why value accrues to whoever owns the end-to-end workflow
- [A system of agents brings service-as-software to life](https://foundationcapital.com/system-of-agents/): Foundation Capital. The thesis behind agents replacing outsourced services

## AI risk, governance and security

The controls that get AI past risk review inside the organisation: model risk, operational resilience, AI security and incident handling.

### Who to follow

- [APRA](https://www.apra.gov.au/): Prudential regulator for banks, insurers and super; its operational risk and tech expectations bind AI use
- [Australian Signals Directorate / ACSC](https://www.cyber.gov.au/): Issues security guidance on AI systems and advisories that feed into board cyber risk
- [Gradient Institute](https://www.gradientinstitute.org/): Local, technical work on AI risk and assurance, often with government and industry
- [AI Safety Newsletter](https://newsletter.safe.ai/): Fortnightly roundup of AI safety research and policy, for risk teams tracking frontier issues
- [Joanne C Klein](https://joannecklein.com/): Detailed posts on labelling, retention and data controls that decide whether Copilot is safe to roll out

### Start with these

- [Voluntary AI Safety Standard](https://www.industry.gov.au/publications/voluntary-ai-safety-standard): Department of Industry, Science and Resources. Australia's ten guardrails for deploying AI, the local reference point for governance
- [AI Risk Management Framework](https://www.nist.gov/itl/ai-risk-management-framework): NIST. The most widely used structure for identifying and managing AI risk
- [ISO/IEC 42001 overview](https://learn.microsoft.com/en-us/compliance/regulatory/offering-iso-42001): Microsoft Learn (compliance documentation). Plain summary of the AI management system standard and how a major vendor certifies to it
- [Operational risk management (CPS 230)](https://www.apra.gov.au/operational-risk-management): APRA. The prudential standard that governs AI vendors and critical operations for regulated entities
- [Advancing accountability in AI](https://oecd.ai/en/accountability): OECD. How to connect risk-management frameworks to accountability across the AI lifecycle
- [Can the finance sector oversee AI innovation while maintaining its rapid progress?](https://oecd.ai/en/wonk/can-the-finance-sector-oversee-ai-innovation-while-maintaining-its-rapid-progress): OECD. Supervisory thinking on AI in financial services
- [Towards a common reporting framework for AI incidents](https://www.oecd.org/en/publications/towards-a-common-reporting-framework-for-ai-incidents_f326d4ac-en.html): OECD. A template for logging and reporting AI incidents consistently
- [Disrupting the first reported AI-orchestrated cyber espionage campaign](https://www.anthropic.com/news/disrupting-AI-espionage): Anthropic. A real incident showing why agent security belongs on the risk register

## Evals and AI quality

Evals are how you prove an AI system works and keep proving it after every model change. The practitioners who made evals a discipline.

### Who to follow

- [Hamel Husain](https://hamel.dev/): The best practical source on error analysis, LLM-as-judge and building evals that matter
- [Eugene Yan](https://eugeneyan.com/): Deeply researched patterns for evals, retrieval and recommendation systems; posts less often but each one lasts
- [Jason Liu (jxnl)](https://jxnl.co/): Concrete playbooks for improving RAG with measurement rather than guesswork
- [Arize AI Blog](https://arize.com/blog/): Frequent material on tracing, evaluation and monitoring LLM apps in production

### Start with these

- [Your AI Product Needs Evals](https://hamel.dev/blog/posts/evals/): Hamel Husain. Shows why evals drive iteration speed and how to start with simple assertions and traces
- [Evals FAQ](https://hamel.dev/blog/posts/evals-faq/): Hamel Husain. Practical answers to the questions every team asks when starting evals
- [Demystifying evals for AI agents](https://www.anthropic.com/engineering/demystifying-evals-for-ai-agents): Anthropic. How to evaluate multi-step agents, not just single answers
- [What We've Learned From A Year of Building with LLMs](https://applied-llms.org/): Yan, Bischof, Frye, Husain, Liu, Shankar. Tactical, operational and strategic lessons from six practitioners in one place
- [A five-step roadmap to closing the AI evaluation gap](https://oecd.ai/en/wonk/a-five-step-roadmap-to-closing-the-ai-evaluation-gap): OECD. The policy view of why evaluation lags deployment, and what to do

## Data foundations for AI

Agents are only as reliable as the data and definitions beneath them. Semantic layers, data contracts, governance and the platforms that make data AI-ready.

### Who to follow

- [Joe Reis](https://joereis.substack.com/): Practical, opinionated commentary on data teams, modelling and the industry
- [benn.substack (Benn Stancil)](https://benn.substack.com/): Sharp, often contrarian essays on what data teams are for and where the market is going
- [Data Products (Chad Sanderson)](https://dataproducts.substack.com/): The main source on data contracts and shifting data quality upstream; posts less often
- [Juan Sequeda](https://juansequeda.substack.com/): Connects data semantics and metadata to making AI work on enterprise data
- [dbt Labs Blog](https://www.getdbt.com/blog): Product direction and practice for analytics engineering and the dbt semantic layer
- [The Analytics Engineering Roundup](https://roundup.getdbt.com/): Weekly curation and commentary on analytics engineering and data plus AI
- [Databricks Blog](https://www.databricks.com/blog): Primary source for lakehouse architecture, governance and data plus AI platform changes
- [Snowflake Blog](https://www.snowflake.com/en/blog/): Primary source for Snowflake platform, Iceberg and data sharing changes
- [Data Engineering Weekly](https://www.dataengineeringweekly.com/): Rounds up the best engineering posts on pipelines, platforms and data quality each week
- [Simon Späti (ssp.sh)](https://www.ssp.sh/): Detailed comparisons of data tools, semantic layers and open table formats
- [Kai Waehner](https://www.kai-waehner.de/): Real enterprise case studies of Kafka, Flink and streaming architectures by industry
- [Jack Vanlightly](https://jack-vanlightly.com/): Rigorous deep dives into how Kafka, Iceberg, Delta and Hudi really behave

### Start with these

- [The Rise of Data Contracts](https://dataproducts.substack.com/p/the-rise-of-data-contracts): Chad Sanderson. Makes the case for agreed schemas between producers and consumers to stop pipeline breakage
- [What is analytics engineering?](https://www.getdbt.com/blog/what-is-analytics-engineering): dbt Labs. Defines the analytics engineer role and the software practices it brings to analytics
- [Data Mesh Principles and Logical Architecture](https://martinfowler.com/articles/data-mesh-principles.html): Zhamak Dehghani. The four principles and the data product idea in their defined form
- [Lakehouse: A New Generation of Open Platforms that Unify Data Warehousing and Advanced Analytics](https://www.cidrdb.org/cidr2021/papers/cidr2021_paper17.pdf): Armbrust, Ghodsi, Xin, Zaharia (CIDR 2021). The paper that defined the lakehouse architecture
- [AI, data governance and privacy](https://www.oecd-ilibrary.org/science-and-technology/ai-data-governance-and-privacy_2476b1a4-en): OECD. Where data governance and privacy law meet AI
- [Context graphs: capturing the why in the age of AI](https://www.linkedin.com/pulse/context-graphs-capturing-why-age-ai-dharmesh-shah-oyyze): Dharmesh Shah. Why agents need the reasons behind data, not just the data

## AgentOps and AI platforms

Running AI as a platform: Copilot and assistant rollouts, governing prompts, skills and agents, shared context, connectors and the cost per seat.

### Who to follow

- [Office 365 for IT Pros](https://office365itpros.com/): Near-daily, candid analysis of Microsoft 365 Copilot features, licensing and admin gaps
- [Microsoft 365 Copilot Blog (Tech Community)](https://techcommunity.microsoft.com/category/microsoft365copilot/blog/microsoft365copilotblog): Feature releases, admin controls and deployment tips straight from the product team
- [Microsoft Foundry Blog](https://devblogs.microsoft.com/foundry/): Hands-on posts on models, agents and evaluation tools in Azure's AI platform
- [Microsoft 365 Roadmap](https://www.microsoft.com/en-us/microsoft-365/roadmap): Item-level feed of upcoming Microsoft 365 and Copilot changes, useful for change management
- [AWS Machine Learning Blog](https://aws.amazon.com/blogs/machine-learning/): Worked examples on Amazon Bedrock and SageMaker, often from named enterprise customers
- [Google Cloud Blog: AI & Machine Learning](https://cloud.google.com/blog/products/ai-machine-learning): Updates and case studies on Vertex AI and Gemini Enterprise
- [Google Workspace Updates](https://workspaceupdates.googleblog.com/): Each Gemini-in-Workspace change and its admin settings, as it ships
- [Salesforce News](https://www.salesforce.com/news/): Agentforce and Einstein announcements, pricing changes and customer stories
- [ServiceNow Blog](https://www.servicenow.com/blogs.html): Now Assist and AI agent updates for IT and service workflows

### Start with these

- [Microsoft 365 Copilot adoption hub](https://adoption.microsoft.com/en-us/copilot/): Microsoft. Official rollout playbooks, scenario libraries and success measures for Copilot
- [Data, privacy and security for Microsoft 365 Copilot](https://learn.microsoft.com/en-us/copilot/microsoft-365/microsoft-365-copilot-privacy): Microsoft Learn. What Copilot can see, where data goes and which controls apply; the first doc risk teams ask for
- [How many products does Microsoft have named Copilot?](https://teybannerman.com/strategy/2026/03/31/how-many-microsoft-copilot-are-there.html): Tey Bannerman. A map of every Copilot, useful before any licensing decision
- [AI adoption in the Cloud Adoption Framework](https://learn.microsoft.com/en-us/azure/cloud-adoption-framework/scenarios/ai/): Microsoft Learn. Step-by-step guidance on strategy, governance and platform setup for AI on Azure
- [Generative AI Lens, AWS Well-Architected Framework](https://docs.aws.amazon.com/wellarchitected/latest/generative-ai-lens/generative-ai-lens.html): Amazon Web Services. Checklist for building reliable, secure and cost-aware generative AI workloads
- [Building A Generative AI Platform](https://huyenchip.com/2024/07/25/genai-platform.html): Chip Huyen. A step-by-step reference architecture from simple calls to routing, guardrails and caching

## AI strategy and economics

Where AI creates value and where it does not: business models, pricing, pilot-to-production economics and the market signals executives watch.

### Who to follow

- [Benedict Evans](https://www.ben-evans.com/): Long-view essays and an annual deck on where AI fits in the wider tech and business cycle
- [Stratechery](https://stratechery.com/): Explains the strategic moves of the big AI vendors you buy from; free posts appear in the public feed
- [Exponential View](https://www.exponentialview.co/): Weekly read on AI adoption, costs and economic impact, with data a board can follow
- [Platforms, AI and the economics of BigTech](https://platforms.substack.com): Sharp analysis of how AI reshapes business models, workflows and competitive position
- [Tomasz Tunguz](https://tomtunguz.com): Short, data-led posts on AI economics, pricing and the software market
- [Clouded Judgement](https://cloudedjudgement.substack.com): Weekly view of software and AI company metrics for anyone building a business case
- [The Leverage](https://gettheleverage.com): Strategy essays on profit and power in AI, from a former Every lead writer
- [a16z](https://www.a16z.news/): Regular data on what enterprises are buying and building with AI, from an investor's seat
- [McKinsey Insights](https://www.mckinsey.com/featured-insights): Source of the widely cited State of AI survey and operating-model research; feed covers all topics
- [Understanding AI](https://www.understandingai.org/): Clear, sober explainers of what AI can and cannot do yet, useful for setting realistic expectations
- [Big Technology](https://www.bigtechnology.com/): Interviews and reporting on how large companies are deploying AI and what it costs them
- [Epoch AI](https://epoch.ai/): Data-backed forecasts on scaling and cost that are useful for planning

### Start with these

- [AI Index Report](https://hai.stanford.edu/ai-index): Stanford HAI. Annual, data-heavy baseline on AI capability, cost, adoption and investment
- [Presentations (including the annual AI deck)](https://www.ben-evans.com/presentations): Benedict Evans. Board-ready framing of how AI shifts the tech industry and where value may land
- [AI horseless carriages](https://koomen.dev/essays/horseless-carriages/): Pete Koomen. Why bolting AI onto old software designs disappoints
- [Slow-burn AI: when augmentation, not automation, is the real threat](https://platforms.substack.com/p/slow-burn-ai-when-augmentation-not): Sangeet Paul Choudary. How augmentation quietly reshapes roles and margins
- [Venture capital investments in AI through 2025](https://oecd.ai/en/ai-publications/venture-capital-investments-in-artificial-intelligence-through-2025): OECD. Where the money in AI is going, by country and sector

## Builders' corner

For AI engineers and deep practitioners: architectures, context engineering, inference, research and the people shipping it. The nerdy end of the hub.

### Who to follow

- [Simon Willison's Weblog](https://simonwillison.net/): The most consistent hands-on log of what coding agents, tool use and new models actually do when you try them
- [The Pragmatic Engineer](https://newsletter.pragmaticengineer.com): The most-read engineering newsletter; tracks how AI tools change real software teams
- [Andrej Karpathy](https://karpathy.bearblog.dev): Clear explanations of how LLMs work and where software is heading
- [Latent Space](https://www.latent.space/): Long interviews with the people building frontier models and AI products
- [AI News by Latent Space](https://latent.space/news/): The fastest way to catch each day's model releases and discussion in one email
- [Interconnects](https://www.interconnects.ai/): Insider-level analysis of RL, post-training and open models
- [Import AI](https://importai.substack.com/): Reads important papers each week and explains why they matter for policy and capability
- [Ahead of AI (Sebastian Raschka)](https://magazine.sebastianraschka.com/): Clear monthly reviews of new LLM architectures and papers, often with code
- [Martin Fowler](https://martinfowler.com/): Sober, engineering-first essays and fragments on using coding agents in real teams
- [Elevate (Addy Osmani)](https://addyo.substack.com/): Practical guidance on working with AI coding agents and keeping code quality up
- [Armin Ronacher](https://lucumr.pocoo.org/): Detailed first-person accounts of building software with coding agents, including what fails
- [Geoffrey Huntley](https://ghuntley.com/): Pushes coding agents to their limits and documents the techniques, such as running agents in loops
- [Hugging Face Blog](https://huggingface.co/blog): Release notes and tutorials on open models, fine-tuning, quantisation and inference
- [vLLM Blog](https://vllm.ai/blog): Where to learn how serving performance, batching and new model support actually work
- [Modal Blog](https://modal.com/blog): Engineering posts with real numbers on GPU inference, latency and cost
- [Together AI Blog](https://www.together.ai/blog): Research and benchmarks on fast inference, speculative decoding and fine-tuning
- [Doug Turnbull](https://softwaredoug.com/): Grounded view of retrieval and search ranking that RAG teams often skip
- [Deep (Learning) Focus (Cameron R. Wolfe)](https://cameronrwolfe.substack.com/): Thorough explanations of post-training, RL and fine-tuning methods with paper references
- [Vicki Boykis](https://vickiboykis.com/): Careful, unhyped writing on embeddings, ML infrastructure and building with models
- [Lil'Log (Lilian Weng)](https://lilianweng.github.io/): Rare but exhaustive literature reviews that become standard references
- [Anthropic Research](https://www.anthropic.com/research): Primary source for interpretability, alignment and Claude capability work
- [OpenAI News](https://openai.com/news/): Primary source for OpenAI model, research and product releases
- [Google DeepMind Blog](https://deepmind.google/blog/): Primary source for Gemini and DeepMind science results
- [Google Research Blog](https://research.google/blog/): Steady stream of applied research across ML systems, health and science

### Start with these

- [Patterns for Building LLM-based Systems & Products](https://eugeneyan.com/writing/llm-patterns/): Eugene Yan. Seven patterns (evals, RAG, caching, guardrails and more) that still map most production systems
- [Introducing Contextual Retrieval](https://www.anthropic.com/research/contextual-retrieval): Anthropic. A measured technique for cutting RAG retrieval failures, with numbers
- [LLM Powered Autonomous Agents](https://lilianweng.github.io/posts/2023-06-23-agent/): Lilian Weng. The canonical breakdown of planning, memory and tool use that most later work builds on
- [How to write a good spec for AI agents](https://addyosmani.com/blog/good-spec/): Addy Osmani. Specs as the main lever on agent output quality
- [The learning loop and LLMs](https://martinfowler.com/articles/llm-learning-loop.html): Martin Fowler. Why developers still need to learn while AI writes code
- [How generative and agentic AI shift concern from technical debt to cognitive debt](https://margaretstorey.com/blog/2026/02/09/cognitive-debt/): Margaret-Anne Storey. The hidden cost when teams stop understanding their own systems
- [The Bitter Lesson](http://www.incompleteideas.net/IncIdeas/BitterLesson.html): Rich Sutton. The short essay explaining why general methods that scale with compute keep winning
- [Scaling Monosemanticity](https://transformer-circuits.pub/2024/scaling-monosemanticity/): Anthropic (Transformer Circuits). Landmark interpretability result showing interpretable features inside a production model
- [Can AI Scaling Continue Through 2030?](https://epoch.ai/publications/can-ai-scaling-continue-through-2030): Epoch AI. Quantifies the power, chip, data and latency limits on training scale

## Australia and New Zealand

Local news, programmes and voices on AI and data across Australia and New Zealand.

### Who to follow

- [iTnews](https://www.itnews.com.au/): Daily reporting on how Australian banks, agencies and large firms run tech and AI projects
- [InnovationAus: Artificial Intelligence](https://www.innovationaus.com/): Close coverage of federal AI policy, procurement and regulation debates
- [The Mandarin](https://www.themandarin.com.au/): How government agencies are adopting and governing AI and data
- [CSIRO's Data61](https://www.csiro.au/en/about/people/business-units/data61): Australia's largest AI research group, including responsible AI engineering work
- [AI Forum New Zealand](https://aiforum.org.nz/): NZ adoption research, policy submissions and member case studies
- [Tech Council of Australia](https://techcouncil.com.au/): Industry positions on AI regulation, skills and investment that shape national policy
- [Memia](https://memia.com/): Weekly ANZ-aware digest of AI and tech developments with local policy commentary
- [The Next Billion Seconds](https://nextbillionseconds.com/): Australian podcast and essays on where AI is heading and what it means for society
- [SmartCompany: Artificial Intelligence (Neural Notes)](https://www.smartcompany.com.au/artificial-intelligence/): How Australian businesses are using AI, with local vendors and data
- [IT Brief Australia](https://itbrief.com.au/): High-volume local vendor, partner and customer AI announcements

### Start with these

- [National AI Plan](https://www.industry.gov.au/publications/national-ai-plan): Department of Industry, Science and Resources. The Australian Government's overall plan for AI capability, adoption and safety
- [Australia: national AI policy initiatives](https://oecd.ai/en/dashboards/countries/Australia): OECD.AI. Every Australian AI policy initiative in one place, compared with other countries
- [Engaging with artificial intelligence](https://www.cyber.gov.au/resources-business-and-government/governance-and-user-education/artificial-intelligence/engaging-with-artificial-intelligence): Australian Signals Directorate and international partners. Security questions to ask before adopting an AI system
- [Public Service AI Framework](https://www.digital.govt.nz/standards-and-guidance/technology-and-architecture/artificial-intelligence/public-service-artificial-intelligence-framework): New Zealand Government (digital.govt.nz). New Zealand's framework for responsible AI use across the public service

## AI policy and regulation

The external rules: laws, regulators, standards and international frameworks, and how one control set can satisfy several regimes.

### Who to follow

- [Office of the Australian Information Commissioner (OAIC)](https://www.oaic.gov.au/news/media-centre): Sets privacy expectations for AI use under the Privacy Act and publishes enforcement decisions
- [Department of Industry, Science and Resources: Artificial intelligence](https://www.industry.gov.au/science-technology-and-innovation/technology/artificial-intelligence): Home of the Voluntary AI Safety Standard, AI adoption guidance and regulatory proposals
- [Digital Transformation Agency](https://www.dta.gov.au): Owns the policy for responsible use of AI in Australian Government agencies
- [National AI Centre](https://www.ai.gov.au/): Publishes practical adoption guidance, tools and events for Australian organisations
- [NIST Information Technology news](https://www.nist.gov/artificial-intelligence): Announces AI risk, agentic AI security and standards work that global frameworks borrow from
- [The EU AI Act Newsletter](https://artificialintelligenceact.substack.com/): Fortnightly summary of EU AI Act implementation, useful for firms with EU exposure or copying its model
- [Luiza's Newsletter](https://www.luizasnewsletter.com/): Frequent, readable analysis of AI regulation, lawsuits and governance practice
- [IAPP](https://iapp.org/news): Daily coverage of AI governance and privacy law worldwide, written for practitioners
- [Human Technology Institute (UTS)](https://www.uts.edu.au/research/centres/human-technology-institute): Australian research on AI governance for boards and AI law reform

### Start with these

- [OECD AI Principles](https://oecd.ai/en/ai-principles): OECD. The intergovernmental baseline most national AI policies trace back to
- [Policy for the responsible use of AI in government](https://www.digital.gov.au/policy/ai/policy): Digital Transformation Agency. Mandatory rules for Commonwealth agencies, and a useful template for regulated firms
- [Guidance on privacy and the use of commercially available AI products](https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/guidance-on-privacy-and-the-use-of-commercially-available-ai-products): OAIC. How the Privacy Act applies when staff use tools like Copilot or ChatGPT with personal data
- [Common guideposts to promote interoperability in AI risk management](https://www.oecd-ilibrary.org/science-and-technology/common-guideposts-to-promote-interoperability-in-ai-risk-management_ba602d18-en): OECD. Compares major AI risk frameworks side by side
- [Bridging frameworks: how HAIP 2.0 supports interoperability with the EU AI Act](https://oecd.ai/en/wonk/bridging-frameworks-how-haip-2-0-supports-interoperability-with-the-eu-ai-act): OECD. How one reporting effort can serve several regimes
- [Intellectual property issues in AI trained on scraped data](https://www.oecd.org/en/publications/intellectual-property-issues-in-artificial-intelligence-trained-on-scraped-data_d5241a23-en.html): OECD. The copyright and data-scraping questions in plain terms

## Work with me

Contract, interim and fractional work in AI adoption, agentic workflow automation,
AI risk and governance, data foundations, and executive AI coaching. Ways to start: a two-week fixed-fee
business case, delivery lead, fractional AI lead, or coaching. Contact: https://www.linkedin.com/in/tatendagatawadata