About
Energetic and passionate technology leader who empowers those around me to do more with…
Articles by Michael
-
Losing Fidelity in LLMs
Losing Fidelity in LLMs
One thing I've noticed people doing and while many tech folks have warned against the approach, I think it's been…
-
Agentic Protocol Timeline
Agentic Protocol Timeline
In the last 20 months, there have been at least 12 different protocols announced. Many have somewhat overlapping…
-
Paving the Cow Path (now with AI)
Paving the Cow Path (now with AI)
I've noticed that current AI initiatives and products seem to fall into a "Pave the Cow Path" trap. That is, we're…
1 Comment -
Designing AI Ecosystems: Moving Away from Generic "Agents"
Designing AI Ecosystems: Moving Away from Generic "Agents"
The term "agentic", much like the early days of "microservices", currently has a semantic overloading problem…
7 Comments -
Agentic is just microservices?
Agentic is just microservices?
Every executive right now is asking their tech leadership for "Agentic AI." There’s a frantic rush to build multi-agent…
-
Agentic Architectures
Agentic Architectures
One thing I've noticed is that a LOT of folks are talking about individual productivity gains from things like claude…
-
Stop Driving Your AI Strategy by the Rear-View Mirror
Stop Driving Your AI Strategy by the Rear-View Mirror
"What part of our business is non-replicable, and how can we use AI to amplify that?" The Rear-View Mirror Fallacy…
-
The Framework Churn: Balancing Cognitive Evolution, Graph Control, and the "Roll Your Own" Hedge
The Framework Churn: Balancing Cognitive Evolution, Graph Control, and the "Roll Your Own" Hedge
1. The Mirage of the Moat The agentic AI market is trapped in a chaotic gold rush.
-
Scale is where AI code generation goes to wait in line
Scale is where AI code generation goes to wait in line
AI coding in the etnerprise is hitting Amdahl’s Law like a concrete wall Everyone is celebrating a 10x boost in writing…
Activity
1K followers
-
Michael Mainguy shared this#AI when you really don't understand why your model drifts.
-
Michael Mainguy shared thisThis is not specific to target (if it's real, these days who knows), but indicative of what I see happens when companies carve up systems into itty bitty pieces, with multiple teams managing/building "pieces" of an overall system without stepping back and thinking about the overall "big picture".... i.e. "what is our goal?"Michael Mainguy shared thisThis picture sums up Target's struggles. My local one is an absolute mess, and this is par for the course of this store. But this is a systemic problem. 1) Why would the computer system print a sales tag ehen the sale prices >= regular price? 2) Why would the employee tasked with putting up the sale stickers not stop and question? 3) How would I be the first to see this before a second employee, or worse, manager? Sakichi Toyoda would be very disappointed and if Target would adapt just a little bit of his philosophy, they would find a much better path.
-
Michael Mainguy shared thisI feel triggered... do NOT do this if I'm in the room... I will a) immediately stop listening and b) spend the rest of the meeting quietly and politely destroying your credibility. (well.. not really...but dang would I LIKE to) #data #visualizationMichael Mainguy shared thisForge Prep students will never commit chart crimes like this I guarantee it h/t RabbiJacob16 on X
-
Michael Mainguy shared thisQuick video of me trying to get claude to refactor my code to include a pluggable database layer: https://lnkd.in/gm5xX-3e #AI #Python
-
Michael Mainguy shared thisAre you confused by what "Agentic" means...you're not alone. My proposal to help reduce confusion and miscommunication. #ai
-
Michael Mainguy shared thisFascinating data! Can't wait for BP to publish their 'Crude Oil Economic Index' showing how much faster cars can move compared to horses. The title should be something more like "Our observations of what claude users are doing"...with a bunch of assumptions about how we that that 'should' correlate with economic output and our assessment of how it might affect the overall economy" (or something like that). Let's stop dressing up vendor opinions as independent research, it's tiring.Michael Mainguy shared thisWe’re publishing our fourth Anthropic Economic Index report, which introduces what we're calling "economic primitives"—foundational metrics on how people actually use AI: task complexity, education level, purpose (work, school, personal), AI autonomy, and success rates. A few findings stand out. AI speeds up complex tasks more than simpler ones: the higher the education level needed to understand a prompt, the more AI reduces how long it takes—even accounting for lower success rates on harder tasks. Productivity gains are real, but uneven. We estimate AI could add 1-2 percentage points to annual US labor productivity growth over the next decade, with the largest gains in complex, knowledge-intensive work. We also see geographic variation in usage—as GDP per capita increases, people use Claude more for work or personal purposes; as it decreases, they're more likely to use it for coursework. And AI is spreading fast. Within the US, if current adoption trends continue, Claude usage could equalize across all 50 states within 5 years—roughly 10x faster than major technologies spread in the 20th century. The clearest takeaway is that AI's impact on global work remains uneven—concentrated in specific countries and occupations, with different effects in each. Read the blog: https://lnkd.in/gdueiz_7. And see the full report here: https://lnkd.in/gtEYK3jyThe Anthropic Economic Index report: New building blocks for understanding AI useThe Anthropic Economic Index report: New building blocks for understanding AI use
-
Michael Mainguy reacted on thisMichael Mainguy reacted on thisForge Prep students will never commit chart crimes like this I guarantee it h/t RabbiJacob16 on X
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Licenses & Certifications
Volunteer Experience
-
-
Youth Soccer Coach
Rockford Raptors Youth Soccer
- Present 12 years
Children
Languages
-
English
Native or bilingual proficiency
-
German
Limited working proficiency
Organizations
-
IEEE Computer Society
-
- Present
Recommendations received
4 people have recommended Michael
Join now to viewOther similar profiles
Explore more posts
-
Starseer
AI Assurance for the Enterprise AI is powering mission-critical workflows—but Shadow AI, compliance gaps, and opaque models put your org at risk. Starseer delivers the visibility, control, and confidence your teams need: > Shadow AI Scanning & Management: Uncover every internal and external AI model—approved or not. > AI Compliance Standards Mapping: Map usage to MITRE ATLAS, NIST AI RMF, ISO 42001 & more, tracking coverage and close gaps with precision. > AI Proxy & Guardrails: Embed real-time policy enforcement into AI workflows. > Model Interpretability: Deep analysis to detect backdoors, jailbreaks, bias, drift, and misaligned agents. With Starseer, organizations can perform a AI Census, gain actionable insights, automated governance, and enterprise-ready guardrails to deploy assured AI. #AIsecurity #ShadowAI #AIassurance #AIcompliance #Cybersecurity
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content