Nathan Clement

Nathan Clement

@njclement.com

Fintech Staff Engineer. AI & Philosophy. Follower of Christ. Be 100% of who you ought to be!

262 followers 999 following124 posts

An article written by one of my smartest friends at Harvard in which he posits that LLMs are either NOT conscious OR only fleetingly so. Not sure if I agree, but this was one of the most pleasant reading journeys Ive had for years! antonesvillagegreen.substack.com/p/the-boltzm...

antonesvillagegreen.substack.comThe Boltzmann Brain doing my shopping;Or, why Co-Pilot isn’t conscious, or so I think and hope...

In case y'all missed it, cars can do kickflips now 🥹

My wife quit her job as a Medical Assistant to become a software developer, and just published her first article :) njclement.com/writing/good...

njclement.comGood Vibes: My Journey Into AI and Vibe CodingHow 14 years in healthcare led to vibe coding, software engineering, and a new way of helping people through AI and technology.

Quick anonymous poll: how are you actually encountering AI these days? Using it daily, hate it passionately, watching it float by on social, somewhere in between? Seven questions, under a minute, no email collected. Would love responses and shares. docs.google.com/forms/d/e/1F...

docs.google.comAI Opinions PollA short, anonymous poll about how people feel about and use AI. No names, email addresses, or other personal information are collected, and every question is optional.

Seems like it has been getting harder and harder to get out into nature in Oregon. Last paddle-board trip we had to get three permits and pay for parking at two places. Thanks Washington for keeping it easy to get outside! Lacamas Lake was beautiful 🤙

Look. Us regular folk will never be able to raise a clone army in time to fight off the Silicon Valley droid army unless we have super powerful AI to help us figure out how to accelerate the growth process.

PDX needs to put an MLB Stadium at Gateway. None of this downtown on the waterfront with no parking and blocking hospital access silliness.

My wife left medical to become a vibe coder. Once in a while I review her code, and am regularly in awe of her velocity and work quality. Her secret? She knows exactly what she wants, has incredible attention to detail, and asks her AI assistant to explain its choices.

Everyone’s watching the AI breakthroughs. The actual war? It’s the silent backroom talent deals, culture curation, contracts, counteroffers, people sliding between teams, while everyone else argues about code.

Trying to supercharge Claude agent workflows. Manual "@" calls are a pain, making real orchestration feel out of reach. Has anyone cracked agent delegation, maybe with clever hooks? Craving thoughts from folks in the trenches. Share for visibility if you feel the same way!

Humans write far sloppier #slop than #LLMs could even dream of generating ✍️

You can just ask for things... How will GenAI change data pipeline engineering? #tech #developer #software #data #ai

Freeway was flooded on my way to imaging appointment today at OHSU! #pdx

The video is too long for Bluesky, but if you're worried about AI making Engineers dumber, check this out!

linkedin.comWorried about Cursor and Windsurf rotting your brain? In this video I… | Nathan ClementWorried about Cursor and Windsurf rotting your brain? In this video I discuss why I believe AI will have a different effect on the cognitive ability of engineers, depending on how they approach AI dr...

The next wave of great software engineers will be philosophers and lawyers

Do LLMs think in a language of their own? Do LLMs plan ahead beyond the next token? Why do models hallucinate? Incredible research by Anthropic just dropped: www.anthropic.com/research/tra...

anthropic.comTracing the thoughts of a large language modelAnthropic's latest interpretability research: a new microscope to understand Claude's internal mechanisms

It just works

Starting April 1, Visa is deprecating old Fraud/Dispute management systems in favor of a new, enhanced program called VAMP > Strong incentives for pre-auth prevention > Changes to fee structure > New tools/resources Details below: merchantfraudjournal.com/visa-acquire... What do people think?

merchantfraudjournal.comVisa Acquirer Monitoring Program (VAMP) Changes 2024: New Rules and Compliance TipsAs part of Visa’s ongoing efforts to streamline and enhance its monitoring programs, effective as of March 31, 2025, Visa

Use case for LLM Agents: TDD for Integration Docs. > Step 1: Create fresh application > Step 2: Point Agent at Docs (maybe an llms.txt?) > Step 3: Ask Agent to integrate with your product ✅ Pass = Integration works Automating this process would be a cool product!

Last week I had a "Black Mirror" thought about a potential negative AI future, so I wrote a short story about it. Curious what folks think :) #shortstory articles.njclement.com/we-chose-the...

articles.njclement.comarticles.njclement.com

Testing out our latest integration with the National Do Not Call Registry 📵 AI summary looking nice as well!

When you just built your first AI model and ask your senior why it won’t generalize…

AI does not make simple mistakes, it makes complex ones.

AI generated code is unlikely to forget a semicolon, but it might miss entire branches of logic in a recursive tree analysis; especially if your interfaces are partially or incorrectly defined.

Keeping these pillars strong isn’t just a technical challenge: it’s a strategic imperative for protecting both your users and your bottom line. Read the full article for explanations and stories! dodgeball-engineering.medium.com/4-critical-f...

dodgeball-engineering.medium.com4 Critical Features When Engineering Your Own Anti-Fraud SystemBuilding an anti-fraud system is like constructing a fortress: it must be carefully designed to withstand both external attacks and…

View full thread

Feature 3: Design for Scalability Key Question: Will my system break if I get a lot of traction quickly? Will I have to redesign?

Feature 4: Design for Flexibility Key Question: What if the world changes? Can I adjust quickly and safely?

Keeping these pillars strong isn’t just a technical challenge: it’s a strategic imperative for protecting both your users and your bottom line. Read the full article for explanations and stories! dodgeball-engineering.medium.com/4-critical-f...

dodgeball-engineering.medium.com4 Critical Features When Engineering Your Own Anti-Fraud SystemBuilding an anti-fraud system is like constructing a fortress: it must be carefully designed to withstand both external attacks and…

Feature 1: Design for Observability Key Question: How do I know what is happening, and how that impacts my business?

Feature 2: Design for Testability Key Question: How do I know a new approach won’t kill my system?

How can you protect a Shopify store from fraud using the Dodgeball App?? Video breakdown 0:00 Introduction 0:31 The Checkpoint Editor 0:42 A Simple Test for Emails 1:45 Email Analysis with AtData 2:52 Test Order on Shopify 3:18 Payment Protection with Sift 4:48 More Capabilities 5:27 Summary

youtu.beProtecting your Shopify Store with DodgeballYouTube video by Dodgeball

Someone should build an app called Mt. Olympus, where top models debate as agents until they reach a mutually agreed solution to complex queries, and you can watch the "distributed" chain of thought live. Too many ideas and too little time to do it myself :P

What should we take away from the DeepSeek drama over the last weekend? 1. Complacency loses 2. Historical success is not a guarantee of future outcomes 3. Rearguard actions (like limiting access to NVIDIA GPUs) are not solution Innovate, don't entrench.

Andrew@Dodgeball 🍁@andrew-at-sea.bsky.social

“Every bubble is unique until it bursts” When a new technology comes about, it's comfortable to think that existing money and hiring an entire MIT class will protect the IBMs of the day. Whereas the real lesson of innovation is that small groups of the smart/hungry/driven always dominate.

I listen to audiobooks every morning. I get physical copies of my favorites. One of these favorites is “Extreme Ownership” by Jocko Willink and Leif Babin This pictured section below struck me today. “Simple”

In software, reducing complexity is often called cleaning. Clean Code doesn’t automatically make conversations with customers or internal teammates simple. Extend your pursuit of simplicity beyond the product, and impact your whole organization.

Curious about using Dodgeball from an Engineering Perspective? Check out our first video tutorial, designed to show engineers everything they might need to know about working with Dodgeball! #software #engineer #dev #fraud #fraudfighter

youtu.beZero to Hero: Everything Engineers need to know about Configuring Dodgeball (Next.js Edition)YouTube video by Dodgeball Engineering

One example of a value-added use of LLMs: At @dodgeball-eng.bsky.social - we are crafting Agentic AI Workflows to augment and normalize information, providing predictable inputs for decisions: 🧵(1/3)

Andrew@Dodgeball 🍁@andrew-at-sea.bsky.social

It isn't AI if it just gets in the way, and using an LLM doesn’t make your software useful. Clippy's just annoying: he doesn’t help, keeps popping up, and forces you to look at him. The real challenge in AI isn't the algorithms: it's figuring out how to make it additive to a working system.

Readings could we export: - Is this a fictional name? - Is this an influencer in industry X? - Is this email likely the primary one for this name? - Is this a business or personal entity? - Does this email look script-generated? Are we seeing many orders to related names/emails? 🧵(2/3)

This opens up new possibilities for fighting fraud unobtrusively, incrementally, and without adding risk. 🧵(end)

Nathan Clement reposted

Andrew@Dodgeball 🍁

@andrew-at-sea.bsky.social

It's always a good day when one of your customers is willing to talk about using your software to save money: www.linkedin.com/posts/mercha...

linkedin.comMRC | Merchant Risk Council on LinkedIn: #paymentsoptimizationIn a rapidly growing subscription business, maintaining a proactive fraud posture is critical to optimizing both cost and effectiveness. Building a scalable…

How do engineers stay relevant in an #ai driven future? Get good at knowing what to build, and describing it in elegant, technical detail

The inaugural article by the Dodgeball Engineering Team is now live! Uncovering Engineering Best Practices for Fraud Better Protection with Less Work Let us know what you think of our lessons learned 🔬 dodgeball-engineering.medium.com/uncovering-e...

dodgeball-engineering.medium.comUncovering Engineering Best Practices for FraudBetter Protection with Less Work

Really excited about an article in the pipeline explaining how to think about fraud in 2025 as an Engineer. Here's a little taste!

Wanting useful things is the new superpower

v0 by @vercel.com could become the only mockup tool engineers need.

It took my poorly drawn art project and transformed it into a working mini-app in just a few prompts.

Playing with a functional preview let me iterate away clunky aspects of my concept before they became baked into requirements. Plus, members of the team can now iterate from a common artifact - far more visceral and informative than screenshots and text explanations.

Easing back into the year with some food engineering 🤣 Our first homemade sourdough and butter turned into some mighty nice breakfast! Credit to @OpenAI for the sourdough recipe and my wife’s Italian heritage for knowing what to do with it🧑‍🍳