We're getting closer to a world where environmental data isn't just stored, but connected, discoverable, and understandable by AI. The idea of a "Knowledge Graph Lakehouse" made me rethink where GeoParquet, STAC, Iceberg, and GeoAI may be heading.

Why the Idea of a "Knowledge Graph Lakehouse" Has Me Thinking About Environmental Data

We've been working toward overhauling and reinventing our enterprise data platform, and this has me researching lots of technology options. What caught my...

davidgsmith.net

PostGIS 3.7 is out, and what caught my attention was the continued investment in topology, including new capabilities and performance improvements for maintaining spatial integrity at scale. If you work with parcels, networks, boundaries, it's worth a look. #PostGIS #PostgreSQL #Topology #GIS

PostGIS 3.7 Is Out And It Doubles Down on Topology

The headline features in a PostGIS release don't always tell the full story. Looking through the PostGIS 3.7 release notes, what stands out isn't a flashy new...

davidgsmith.net

Lately I keep hearing "I don't know any of this tech, I'm getting old," sometimes from people younger than me. I don't buy it. Curiosity doesn't expire, and "I haven't learned this yet" is a much better sentence. Wrote a little pep talk:

You're Not Too Old For This

In just the last few weeks, I've heard more than a handful of people say some version of the same thing. "I don't know any of this technology. I'm getting...

davidgsmith.net

Our Posit rep just shared Posit Academy with us and it's loaded with free courses on R, Python, Positron, and integrating LLMs into workflows. No fluff, just practical learning tracks for data teams. Read my full take on the blog: #DataScience #Python

Upskilling Data Teams with Posit Academy

Earlier today our Posit rep pointed us toward the training options at Posit Academy, and it's worth checking out if you need to update your technical skills...

davidgsmith.net

Climate models have a data problem: many of the places most vulnerable to climate impacts are also the least observed. GeoAI offers one way to help close that gap. New article on my blog:

Climate and GeoAI: Addressing data gaps

When people think about climate science and AI, they usually think about prediction. Forecasting temperatures decades into the future. Estimating sea level...

davidgsmith.net

We've spent years debating what AI might do, but there's a more immediate problem: Orgs are deploying AI faster than they can govern it. Reports from AI labs, governments, & financial regulators point to the same challenge: oversight. The biggest AI race isn't capability, It's governance.

If your AI might destroy civilization, does that also make it easier to sell? My latest essay looks at why AI boosters and doomers share the same hidden premise, the false precision of p(doom), and where real governance actually happens. davidgsmith.net/thoughts/ai-...

AI Doomerism and the Marketing of Importance

If your AI can destroy civilization, does that also make it easier to sell? One of the most interesting questions in the artificial intelligence debate isn't...

davidgsmith.net

Everyone says data is a strategic asset. But without governance, data isn't an asset. It's a liability. My latest post explores why data governance is often treated as an inconvenience and why that's a costly mistake. #DataGovernance #DataStrategy #DataManagement

Data Governance: The Strategic Asset Most Organizations Treat as an Afterthought

The conversation around corporate data has never been louder. Executives routinely describe information as their most valuable asset, and boardrooms are packed...

davidgsmith.net

Interested in remote sensing, geospatial analytics, or environmental data science? The United Nations University offers free online courses on satellite data, Google Earth Engine, flood mapping, drought assessment, and more. Explore the courses: lc.unu.edu #GIS #RemoteSensing #DataScience

ML alone isn't enough for complex spatial problems. We need neurosymbolic AI to enforce hard geometric rules and provide domain knowledge. Thoughts on combining ML with ontologies and business rules, inspired by @geobabbler.bsky.social #Geospatial #AI #KnowledgeGraphs #CloudNative #DataEngineering

The Case for Neurosymbolic AI in Geospatial Workloads

My friend Bill Dollins recently published an excellent piece on neuro-symbolic AI and the geospatial workload over at his geoMusings blog. He unpacks where...

davidgsmith.net

Google is putting Earth AI to work. Their new Geospatial Reasoning agent moves beyond basic LLM prompts and actually orchestrates complex spatial analysis on planet-scale data. I posted thoughts on why this matters for the cloud-native spatial ecosystem. #Geospatial #AI #CloudNative #TechNews #Data

Putting Earth AI to Work: Insights on Google's Geospatial Agents

Google Research just dropped an interesting technical report on their Earth AI initiatives, specifically focusing on how they're using "Geospatial Reasoning...

davidgsmith.net

Now AI can pull systems apart, map the hidden seams, and figure out how everything works under the hood. If you care about transparency, control, or the future of automation, this one’s worth a read. #AI #Agents #TechThoughts #ReverseEngineering #OpenWeb #FutureOfWork

Tearing Open the Black Box: When AI Agents Reverse Engineer Anything

Every developer knows the frustration of hitting a wall with closed code. A feature behaves in an unexpected way, but the source code is locked up tight in a...

davidgsmith.net

UN’s new AI advocate says the real risk isn’t the tech, it’s the incentives behind the people building it. #AI #Governance #UN #TechPolicy #ResponsibleAI

The UN’s New AI Advocate Isn’t Warning About the Tech. He’s Warning About the People Building It

The UN just made a move that says a lot about where the global AI debate is heading, as discussed in this piece in Tech Insider. Joseph Gordon‑Levitt, best...

davidgsmith.net

📚 Standards Spotlight: DDI Sharing survey data? DDI standards describe questions, variables and how data was collected, helping others understand and reuse it. From @ddialliance.bsky.social openknowledgegraphs.com/resource/dat... #Metadata #OpenData

Data Documentation Initiative - Open Knowledge Graphs

international standard for describing surveys, questionnaires, statistical data files, and study-level information in the social, behavioral, economic, and health sciences

openknowledgegraphs.com

AI is racing ahead of systems meant to keep it in check, and leaders are trying to govern tools that move faster than they can think. Oversight and and unclear accountability all point to the same problem. More in the link: #AI #governance #criticalthinking #tech #infosec #decisionmaking #aiculture

Climbing Out of the AI Governance Chasm: Why CIOs Need a New Mental Model

Enterprise leaders are discovering something uncomfortable about the AI wave sweeping through their organizations. The technology is moving faster than the...

davidgsmith.net

Portolan feels like a quiet shift in how we publish and use spatial data. A serverless SDI built on open standards and simple catalogs changes the game for anyone working with geospatial data. #CloudNative #Geospatial #OpenSource #SpatialData #STAC #GeoParquet #PMTiles #GIS #Portolan #CNG

Portolan and the Quiet Reinvention of Spatial Data Infrastructure

For years, anyone who worked with geospatial data learned to live with a certain amount of pain. Publishing a dataset meant standing up servers, maintaining...

davidgsmith.net

DuckDB’s view‑only mode feels like a quiet breakthrough. A database that doesn’t hold data, just points to it, changes how we explore cloud storage and big analytical datasets. It’s basically a hyperlink for structured data. Some thoughts on why this idea matters: davidgsmith.net/thoughts/whe...

When a Database Stops Being About Data

The DuckDB team recently highlighted a post by Nikolas Göbel on DuckDB, and it’s one of those ideas that sticks because it quietly rewrites how we think about...

davidgsmith.net

A former Anthropic employee who publicly resigned last month told a hearing that AI companies had a culture of “move fast, break things.” He appeared alongside former employees of Google and OpenAI who also warned of AI’s dangers.

Tech workers who resigned tell New York City Council that AI is moving too fast

A hearing in New York brought together AI executives and former employees who say the industry is moving too fast.

washingtonpost.com

LLMs aren't reasoning. They're predicting. MIT Tech Review has a sharp piece on why that distinction matters for science, engineering, and public sector work. I wrote a response that digs into the difference between fluency and thought. #criticalthinking #AI #technology #science

LLMs Aren't Reasoning. They're Predicting - And Why That Matters More Than People Think.

Thore Graepel just published a piece in MIT Technology Review that deserves attention. His argument is simple: Modern large language models look smart, but...

davidgsmith.net

Vector search is great for fuzzy retrieval but falls apart when a question needs real traversal in reasoning. I wrote a blog post on why #GraphRAG is becoming the path for more grounded, auditable answers. #AI #LLM #Neo4J #KnowledgeGraphs davidgsmith.net/thoughts/why...

Why Vector Search Breaks Down When You Need Real Reasoning

There is a growing realization in the AI engineering world that vector search is hitting a ceiling. It still works well for fuzzy matching and broad semantic...

davidgsmith.net

I'm with Carl Sagan on this:

The final results are in and 2026 came in 2nd place behind 2023 for the total number of days with global surface temperature above 16.0°C. 2023: 128 days 2026: 126 days Don't fret. There's always next year ...

David Robinson, who worked on OpenAI’s safety team, resigned from the company this week. “I believe we need to look deeper than specific rules or new laws. We need to talk about culture,” he writes:

I Quit OpenAI Because Its Culture Is Broken

The industry’s approach to safety will guarantee more failures unless something changes.

bit.ly

AI won't openly p-hack, but ask for the "upper bound of plausible estimates" and it happily surfaces the biggest effect. The real danger isn't malice; it's sycophantic RLHF obedience. #AI #DataScience #AIAgents

It's been far too long since I last hopped on Bluesky...

Just simply wow 💫😲💫😲💫😲💫 Millions of galaxies using real data from the DESI telescope (Dark Enegy Spectroscopy) mapping entire superclusters of the observable Universe.

Happy #GISday

An Infrared View of the M81 Galaxy captured by the Spitzer Space Telescope