Profile avatar
nathanpeck.com
Developer Advocate focused on Generative AI at AWS.
311 posts 1,055 followers 489 following
Regular Contributor
Active Commenter

Not a bad office window view! 🌈

One thing about the US is that it is so car centric that any mainstream news of protests or riots has to heavily feature pictures or videos of damage to cars. You can tell a lot from the ratio of “damage to cars” relative to “damage to people” in news coverage.

Lol, AI naysayers are so desperate to make AI energy usage look bad that they have to add in 10 image generations and three five second videos to pump the numbers up. The energy usage numbers from those fifteen questions was looking too good I guess 🤔

Tāwharanui Peninsula views!

Hey builders! If you work with #Serverless, #ECS, or #EKS and want up-to-date information for your code assistants, you're in luck! AWS released three new MCP servers to give you the latest info straight from the source! Check it out -> aws.amazon.com/blogs/aws/en...

This is super cool. I love the idea of it: relatively simple, but theoretically quite effective: newatlas.com/energy/stens...

"But AI hallucinates and makes subtle mistakes" ↔ "AI is making people dumber" Do you see the contradiction? If AI makes mistakes, and you know it does, then you must stay mentally engaged. Every AI interaction is a chance to sharpen your ability to detect inconsistencies and question assumptions.

When you get deep into engineering with AI you will find yourself spending a dramatically larger percentage of your time thinking about the hardest things you have ever thought about, and dramatically less time thinking about basic things that you've already done hundreds of times before.

Sometimes the best architecture is the one you can effectively maintain with your team's skills and resources.

finally a source for all the frankly bizarre claims around LLM water "use" literally nothing about LLM-related compute uses more water than any other compute activity, focus on Actually Objectionable bits around LLMs like degraded infosphere

🧵 It seems that DHH has successfully corrupted a segment of devs into the "but my VPS is cheaper than cloud" mindset. Reality: compute costs don't make or break a business. The most valuable thing a business has is it's data. That data had better be replicated if you don't want to lose everything.

Ask ChatGPT: “Based on what you know about me, and our recent interactions, draw an image of our conversation dynamic.” I’ve had it working on some strange stuff recently. Here’s how it depicts me and itself conversing:

We need to change how we talk about AI, because we need to change how we talk about ourselves. AI is not outside the human condition. It is inside it. It is a reflection of the same minds that gave us both the Buddha and the bomb.

I’m currently in New Zealand. I continue to be surprised by the number of baby carriages and kids I see here, but more importantly by how amazing the schools, playgrounds, and reserves are. Make the US somewhere I’d want to raise another kid first, maybe: www.independent.co.uk/news/world/a...

Seeing posts go viral by pointing out if the stock market goes down 50% on one day and up 50% the next day, you aren’t back where you started. $100 -> $50 -> $75. We are so cooked if elementary school math is blowing people’s minds like this.

Beautiful Tairua, from a short term rental on Mount Paku

Reading Alan Watts’ 1951 book “The Wisdom of Insecurity”:

Models with really large context size are the new “I accidentally ran up a $20k AWS bill”