Here’s a roundup of interesting startup links I came across today:
Connecting the wisdom of Yoda to the building of startup communities | by Mitch Daugherty | Built Oregon | Medium
Most startup communities are open, willing, and collaborative by nature. No different Oregon is. There is a passion amongst the founders and support organizations to work together and accelerate their local and regional economies. But creating change, and being able to determine how best to engage and accelerate, is something that requires several commitments.
Stop Taking a Helicopter to the Grocery Store — Antfly Blog
LLMs are general-purpose reasoning engines. They’re extraordinary at understanding context, generating text, and handling ambiguous instructions. But most of the steps in an AI pipeline aren’t ambiguous. Chunking is chunking. Embedding is embedding. Reranking is reranking. These are well-defined, narrow tasks — and there are models specifically trained to do each one faster, cheaper, and often better than a general-purpose LLM.
Fear and Loathing in the Software Community – Elia Insider
Yes, our world is changing but the need for software development doesn’t end here. First things first, the act of software development is the act of describing a product in unambiguous detail*. Most people can’t do this. Being extremely detail oriented is a trait found in few professions, software development being one of them.
Burn Bright, Not Out | Feld Thoughts
May is Mental Health Awareness Month. I’ve been writing about my own depression and mental health on this blog for over a decade. The stigma around founder mental health is still the core problem. 72% of founders say the entrepreneurial journey has negatively affected their mental health. 81% don’t talk about it.
Pitch Latino 2026 Q&A
Join us for a brief overview of Pitch Latino, a signature entrepreneurial showcase and pitch competition produced by Latino Founders. Designed to celebrate Latino entrepreneurship, innovation, and community impact, Pitch Latino provides founders with visibility, mentorship, and access to powerful networks and resources.
ZeroCVEs is not a Mythos – Minimus
The traditional security playbook revolves around trying to perfectly prioritize that 10%, filtering by severity, active exploits, and reachability. But when AI drastically shrinks the window between vulnerability disclosure and active exploitation, prioritizing a never-ending backlog is a losing battle.
We’re Still #1 in Semifactual Superlatives
Nobody else can claim to be America’s eighth-best metro area for working moms.
Oregon group tasked with creating new transportation solution meets for first time • Oregon Capital Chronicle
“We all depend on a reliable transportation system, and frankly, our system here in Oregon is showing its age and we have work to do,” Kotek told the group at its first Friday meeting at transportation department headquarters in Salem.”I believe we’re losing out on economic development in our state because we don’t have the consistency and certainty that we need.”
How did ‘large’ language models get that way? The role of Transformers and Pretraining in GPT – LessWrong 2.0 viewer
This is part 1 of a series about LLM architecture and some implications, past and future, for reasoning. Part 1 is ‘how we got here’ — what was so impactful about the transformer architecture for LLMs. Some readers may prefer to skip this. Part 2 will point at an unexpected benefit — the surprising explainability of contemporary AI reasoning — and why new trends might erode that. It’s a novel point as far as I know.
🎙️Hugging Face’s Clem Delangue: Stop Comparing Engines to Cars
One of Clem’s clearest points is that comparing open weights to closed APIs is like comparing an engine to a car. And that’s why the question “are open-source models catching up?” doesn’t even matter in the same way. Behind an API, there are tools, harnesses, routing, and sometimes several models. So when people say open models are “behind,” the real question is: behind what system, for what task, and at what cost?
Infrastructure as a Service Builds the Most Defensible Economics
That structural difference applies as much to Stripe, Twilio, and Plaid as it does to Amazon’s server farms. These companies don’t sell to end users; they sell to the companies building products for end users. That’s what makes them infrastructure rather than software, and it’s what’s become a $14 trillion layer of the internet that most founders are either standing on without realizing it or accidentally trying to rebuild from scratch.
Weekly Review: Training the Apprentice – by Sam Keen
Happy May the Fourth, and welcome back to Altered Craft’s weekly AI review for developers. Thanks for spending some of your week with /AC. One thread runs through this edition: the engineer’s role is shifting from writing code to training the apprentice. Karpathy names the discipline, Parsons argues senior engineers should move from approver to trainer, and pieces on prompts as artifacts, skill packs, and harness design all sketch what that practice looks like. The model is the apprentice, and the craft is the training.
How AI Is Changing the Network(s) – On my Om
Over the past few weeks, with help from old friends in the networking and infrastructure world, I have managed to put together a 5,000-word overview of the changes to the network. I look at the physical pipes, the shifting demand profile, and who really owns this new internet.
America’s Electricity Gap – by Joseph Politano
America has entered another era of electrification—US power consumption has risen more in the last two years than over the previous 15 combined. Yet this hasn’t been nearly enough to meet demand, with electricity prices also rising more over the last 4 years than the prior 14. Demand from AI, heavy industry, heating, and transport is driving load growth but also outstripping growth in power infrastructure—leaving America with an increasingly large electricity gap that’s driving up prices.
AddyOsmani.com – Agent Skills
The default behaviour of any AI coding agent is to take the shortest path to “done.” Ask for a feature and it writes the feature. It does not ask whether you have a spec, write a test before the implementation, consider whether the change crosses a trust boundary, or check what the PR will look like to a reviewer. It produces code, declares victory, and moves on. This is the same failure mode every senior engineer has spent their career learning to avoid.
2026 Build Challenge: AI for a Better PDX (Portland Startup Week) · Luma
AI Portland Build Challenge is a one‑day civic‑focused hackathon where teams build AI-powered projects to make life in Portland better, culminating in a 3 pm showcase in front of our panel of judges.
Portland Startup Week: Why Starting Small Just Became The Biggest Edge, Tue, May 12, 2026, 1:30 PM | Meetup
A year ago, Geoff Jensen was at Portland Startup Week trying to figure out the same thing a lot of founders are: find a co-founder, join a team, or raise money? He won the hackathon, built a great team — and within months realized the team was too big, the fundraising timeline was too slow, and the features they were building were already being absorbed by frontier models.
I haven’t seen you testify yet. (Budget hearings start soon)
I see a lot of familiar faces in City Council public testimony. No matter what the subject, the same lobbyists and special interests show up in the same T-shirts. Their voices are heard by legislators… but what about yours?
