@SteefJan

Cloud Editor @InfoQ, Microsoft Azure MVP, Crew Member @AzureLowLands, Traveller, Runner, #Seahawks, #SeaUsRise, #SeaKraken, Metal Music, Opinions are my own!

 Ede, Gelderland, Netherlands
sjwiggers.com/
Joined May 2008
  • Tweets13.7K
  • Following921
  • Followers2.7K
914Media
628 of 628 Photoshop menus implemented. ~45% ready for real work. The dev behind the AI-built Adobe clones walked back “parity in a month.” His own roadmap now says 2,300-4,100 more agent hours — and its PSDs still don’t open reliably in Photoshop. infoq.com/news/2026/10/photo… #ai

Developer Behind the AI-Built Adobe Clones Says They Are “Not Anywhere Close to Ready

A developer used Claude Opus 5.5 to build open-source replacements for seven Adobe applications, promoted them across Reddit with a projection of roughly 100% parity within a month, then told Hacker...

infoq.com
36
Four rules for the MCP layer, built on Azure Functions. Three hold. The fourth (readOnlyHint/destructiveHint) never reaches the wire: the Functions MCP trigger has no field for it. Access is not authority; the API decides. sjwiggers.com/2026/10/22/azu… #azurefunctions #mcp

Access Is Not Authority: Four Rules for the MCP Layer on Azure Functions - Cloud Perspectives

Four rules for the MCP layer, built as an Azure Functions MCP server over a Functions API: own token, task tools, token cost, and the annotation gap.

sjwiggers.com
1
49
Logic Apps Automation agents run code two ways: a JavaScript Code Interpreter, or a sandbox with your repo cloned in. I built a contract reviewer on the sandbox. Correct verdict in 33 s, after 4 failures. Loop count defaults to 5,000. sjwiggers.com/2026/10/20/log… #azure #logicapps

Code Interpreter or Sandbox: Two Ways a Logic Apps Automation Agent Runs Code - Cloud Perspectives

Logic Apps Automation gives agents two ways to run code. I built a contract-review agent in a sandbox: what worked, what failed, and what it costs.

sjwiggers.com
1
1
120
MCP vs RAG vs Skills is the wrong fight. Connection, knowledge, procedure. None decides what the agent may do. Built all three into one Azure agent; run one was right because the model did the math itself and skipped the script. Governance gap. sjwiggers.com/2026/10/16/mcp… #mcp

MCP vs RAG vs Skills Is the Wrong Fight. Governance Is the Real One - Cloud Perspectives

MCP, RAG and Skills are three kinds of context, not rivals. One Azure agent uses all three, and the post shows what governs each layer.

sjwiggers.com
1
1
96
Cloudflare’s new OHTTP Gateway refuses requests that arrive via Cloudflare. On purpose. Relay and gateway must be different parties, or one company sees both who you are and what you asked. The rejection is the feature. infoq.com/news/2026/10/cloud… #ohttp #privacy

Cloudflare’s OHTTP Gateway Refuses to Decrypt Requests That Arrive through Cloudflare

Cloudflare has opened a closed beta of its OHTTP Gateway, a managed endpoint that decrypts Oblivious HTTP requests for an origin. The design choice drawing attention is a refusal: the Gateway will...

infoq.com
91
Maia 300: one Ethernet fabric, KV state moves prefill→decode without leaving it, 288 GB HBM4 per chip. Then the headline: 40% more tokens per megawatt. Tokens per watt isn’t a rate, and nobody names the GPU or the context length. infoq.com/news/2026/10/azure…

Microsoft Commits Maia to Multi-Gigawatt Inference, and Its Efficiency Metric Draws Questions

Microsoft will scale its Maia accelerator family to multi-gigawatt deployments from 2027, making it the primary inference system for its own products. Maia 300 collapses scale-up and scale-out onto...

infoq.com
88
Logic Apps Automation pricing is out. The free grant is worth roughly $21 a month. The environment is free for 60 days, then costs about $30.66 a month, running or idle. Put day 61 in your calendar. sjwiggers.com/2026/10/10/log… #azure #logicapps

Logic Apps Automation Pricing: The Free Grant and What It Changes - Cloud Perspectives

Microsoft announced Logic Apps Automation pricing and a monthly free grant. Here is what the numbers mean, the fine print, and where it does not fit.

sjwiggers.com
1
1
105
Revoking a Foundry agent's data role took 60 s to bite. A kill switch at the APIM gateway took 7.8 s. Default trace sampling quietly dropped 2 of 23 audit steps. Five governance questions, five drills, measured: sjwiggers.com/2026/10/08/fou… #microsoftfoundry #aigovernance

Five Questions, Five Drills: Testing Whether a Foundry Agent Stays Inside Its Bounds - Cloud...

Five questions decide whether an enterprise agent reaches production. I turned them into drills on Microsoft Foundry and measured what holds.

sjwiggers.com
2
1
3
183
20 identical payment requests. 1 idempotency key. 16 charges. All 20 callers got a 201. Twelve API design patterns on Azure, and the one where a green test suite tells you the least. sjwiggers.com/2026/10/07/api… #azure #apidesign

API Design Patterns with Azure - Cloud Perspectives

Twelve API design patterns on Azure, measured. Idempotency keys implemented four ways: three pass the usual test and double charge in production.

sjwiggers.com
3
97
Picking Azure messaging? Start with the payload, not the product. sjwiggers.com/2026/10/03/azu… #azure #servicebus @egrootenboer

Messages, Events or Streams: Choosing Azure Messaging by What You Send - Cloud Perspectives

Azure messaging services compared by what you send: when to use Service Bus, Event Grid, Storage Queues or Event Hubs, with a deployable .NET sample.

sjwiggers.com
1
2
146
Every real-time connection in Azure hits a ~4-minute ceiling. App Service stops at 230s, Front Door at 240s, same load balancer underneath. Six patterns, one limit. Plan the reconnect. sjwiggers.com/2026/10/01/rea… #azure #realtime

Six Real-Time Patterns in Azure: What the Diagram Leaves Out - Cloud Perspectives

Six real-time patterns in Azure, from polling to gRPC streaming, plus the gateway and scaling behavior that breaks them in production.

sjwiggers.com
1
121
Ran Microsoft's Cosmos DB Agent Kit (100+ best-practice rules) against my own sample repo. Caught a real bug: mock embeddings never normalized to unit length, silently breaking cosine-similarity scores. sjwiggers.com/2026/09/29/cos… #agenticai

Cosmos DB Agent Kit: Auditing My Own Sample Repo - Cloud Perspectives

Six posts of hand-checked rules, now checked by a rulebook. I ran the Cosmos DB Agent Kit against my own sample repo — here's what it caught.

sjwiggers.com
1
93
Apple Reference Image signs pixel data inside the iPhone 18 Pro sensor, then develops it in Private Cloud Compute under Apple's signature. Trust moves from C2PA's editing chain to Apple. Verification stays on Apple platforms. infoq.com/news/2026/09/apple…

Apple Reference Image Signs Photos at the Sensor, Moving Provenance Trust away from C2PA

Apple has published the design of Reference Image, an iPhone 18 Pro camera mode that signs pixel data at the sensor and develops it in Private Cloud Compute under an Apple signature. Developers on...

infoq.com
86
A "10 per second" rate limit let 19 through in 200 ms. Add a second replica and every in-process limiter doubled. I load-tested six rate limiting algorithms on Azure, so you don't have to guess. sjwiggers.com/2026/09/24/rat… #ratelimiting #azure

Rate Limiting Algorithms on Azure: 6 Compared - Cloud Perspectives

Six rate limiting algorithms on Azure, measured with real load tests: fixed window, sliding log, token bucket, leaky bucket, and more.

sjwiggers.com
95
1,727 price meters for Foundry Models in one Azure region. None carries a model id. So I built a ledger that joins the ARM catalog, Retail Prices and your own deployments: cost per 1M tokens, retirement dates, what breaks. sjwiggers.com/2026/09/22/fou… #microsoftfoundry #azure

Foundry Model Ledger: Which Model, at What Price, Until When, and Which of Mine - Cloud Perspectives

Foundry Model Ledger shows what a Foundry model costs per 1M tokens in your currency, when it retires, where else it runs, and which of your deployments are affected.

sjwiggers.com
1
99
AWS says it cannot restore data held exclusively in a damaged UAE availability zone, or exclusively in its Bahrain region, after strikes during the conflict with Iran. Multi-AZ and S3 durability are regional properties. infoq.com/news/2026/09/aws-m…

AWS Cannot Restore Data Held Only in Damaged Middle East Availability Zones

AWS has told customers it cannot restore resources and data hosted exclusively in the mec1-az2 availability zone in the UAE, or exclusively in the Bahrain region, after damage during the conflict...

infoq.com
1
131
Same user, session revoked. CAE-capable client lost access in 4 seconds. The other kept going for 65 minutes. Six deployable samples measuring what the Entra ID architecture diagram compresses: sjwiggers.com/2026/09/20/mic… #entraid #conditionalaccess

Microsoft Entra ID under the hood - Cloud Perspectives

Microsoft Entra ID under the hood: token lifetimes, CAE revocation latency and PIM payload, measured on a real tenant with six deployable samples.

sjwiggers.com
104
Real production case: an account at 100% RU utilization, throttling under load. Cause wasn't throughput — one partition absorbing 80% of traffic. Fixing the model dropped utilization to 20-35%. sjwiggers.com/2026/09/17/cos…

Cosmos DB Agent Memory Cost: Caching, RU Drivers, and a Pitfalls Roundup - Cloud Perspectives

Cost only matters next to the value it unlocks. A practical guide to Cosmos DB agent memory cost: semantic caching, RU drivers, and pitfalls.

sjwiggers.com
1
1
91
GPT-6 Astra is the first model OpenAI classifies as Critical for cybersecurity under its Preparedness Framework. In expert-led testing it found unknown vulnerabilities in a browser and an OS kernel and built working exploits. infoq.com/news/2026/09/gpt-6…

GPT-6 Astra is the First Model OpenAI Classifies as Critical for Cybersecurity

OpenAI has classified GPT-6 Astra at the Critical cybersecurity threshold under its Preparedness Framework, a first. In expert-led testing the model found previously unknown vulnerabilities in a...

infoq.com
85
Foundry Agent Service BYO thread storage: Cosmos DB needs 3,000 RU/s minimum or you get CapabilityHostProvisioningFailed. Classic and New runtimes also write to different containers entirely. sjwiggers.com/2026/09/16/cos… #agenticai

Cosmos DB Foundry Agent Service: Bring-Your-Own Thread Storage - Cloud Perspectives

Default agent storage is opaque. A practical guide to Cosmos DB Foundry Agent Service — the enterprise_memory database, RU sizing, and roles.

sjwiggers.com
1
2
105
Load more