Sign in

John Berryman

@jnbrymn.bsky.social
492 followers 390 following 80 posts

LLM consulting @ arcturus-labs.com Author x2: amzn.to/3TXmDHk, amzn.to/3zKIxGG Formerly Eventbrite, GitHub (code search and Copilot)

PostsRepliesMedia
John Berryman @jnbrymn.bsky.social · 22/09/2026
Will OpenAI eat Jev's lunch? Jev's classifier is built on a conventional LLM. OpenAI already knows this trick and uses it. If a more general classifier was folded inside a frontier model, it could provide safe tools, cheap routing, and more. The moat is data. arcturus-labs.com/blog/2026/09...
arcturus-labs.com
Will OpenAI Eat Jev's Lunch?
TypeSafe's Jev is a genuine breakthrough – snap judgments with calibrated probabilities instead of generated text. My bet is OpenAI is already figuring out how to copy it, and then embed it inside its...
110
John Berryman @jnbrymn.bsky.social · 18/09/2026
Everyone's talking about TypeSafe's Jev - decisions instead of text generation. I dug into how it might work, what to build with it, and whether the big labs eat its lunch: arcturus-labs.com/blog/2026/09/16/t…
Blog post hero image
120
John Berryman @jnbrymn.bsky.social · 16/07/2026
I'm super excited to be speaking at O'Reilly's July 23 Superstream. I'll talk about freeing your agent from the terminal and using it as your anything assistant. Be sure to check it out! lnkd.in/eE3ca5c2?
lnkd.in
LinkedIn
This link will take you to a page that’s not on LinkedIn
020
John Berryman @jnbrymn.bsky.social · 19/06/2026
On last night's "Show Us Your (Agent) Skills Episode 5" I demo'd Rook, my personal agent that follows me _anywhere_ and gains the skills necessary to manipulate new environments we encounter. Video: youtube.com/watch?v=6zju... Show notes (w/ timestamps linked to video): github.com/hugobowne/sh...
youtube.com
Show Us Your (Agent) Skills Episode 5 - w/ John Berryman, Isaac Flath, & Matt Palmer
YouTube video by Vanishing Gradients
050
John Berryman @jnbrymn.bsky.social · 16/06/2026
On Thursday I'm presenting a fun project I've been working on. I'm building "Rook", an agent that follows me anywhere (websites, native applications, physical locations). And wherever it goes it gains the skills of it's environment. I have plans for the beast and I need your help! luma.com/0t8kiodw
luma.com
Show Us Your (Agent) Skills Ep. 05 · Luma
What are people at the top of the game building with AI agents and how are they doing it? Are they Claudemaxxing with 8 terminals open at once? Or…
020
John Berryman @jnbrymn.bsky.social · 09/06/2026
Models have gotten good at implementing a mechanical workflow as a skill. However, implementing a workflow that makes decisions based upon "taste" is going to continue to be challenging. How do you all deal with matters of "taste" in AI workflows?
010
John Berryman @jnbrymn.bsky.social · 09/06/2026
I see your emdash :P
000
John Berryman @jnbrymn.bsky.social · 01/06/2026
I built an agent skill that edits video - ums, silences. Recorded a walkthrough on how it works; the skill edited that recording! Post shows setup, edited vs raw, and the skill to fork. arcturus-labs.com/blog/2026/05/31/m…
Chrome robotic hands in an Escher-style loop, each drawing the other from a video frame
011
John Berryman @jnbrymn.bsky.social · 28/05/2026
You don't need to rip out your search stack to add agentic AI to it. A thin AI layer on top of what you have is enough. I mapped out 4 levels of adoption with a live demo at each stage. arcturus-labs.com/blog/2026/01/18/i…
Diagram showing the evolution from traditional keyword search to a full conversational AI search assistant
100
John Berryman @jnbrymn.bsky.social · 03/05/2026
Twin Sun's code review bot approves 70% of PRs on its own. No human sees that code until it's merged. Two principles that make a dev factory work - and neither is 'just be generic.' arcturus-labs.com/blog/2026/05/03/t…
The Dark Factory: How Twin Sun Automated Their Entire Dev Pipeline
130
John Berryman @jnbrymn.bsky.social · 30/04/2026
www.youtube.com/watch?v=kkvi...
youtube.com
The Dark Factory: How Twin Sun Automated Their Entire Dev Pipeline
YouTube video by Arcturus Labs
000
John Berryman @jnbrymn.bsky.social · 29/04/2026
AI product development evolves by learning from prior eras: prompt tricks -> chat/tool use -> workflows -> agents. I mapped the arc from 2022 to now: arcturus-labs.com/blog/2026/03/22/t… Which lesson do we still underuse?
Evolution of AI product eras from early completion models to agentic runtimes
130
John Berryman @jnbrymn.bsky.social · 27/04/2026
We need to Unharness our agents from the IDE. "Agent harness" frames them as coding tools – but agents are the building blocks of the future AI products, carrying context, memory, skills, and interfaces across whatever software we move through. arcturus-labs.com/blog/2026/04...
arcturus-labs.com
Unharnessed Agents Power the Future of AI Products
The next era of AI products will not be built around individual chatbots or coding assistants. It will be built around agents that carry context, tools, skills, and interfaces across every part of dig...
1102
John Berryman @jnbrymn.bsky.social · 19/01/2026
E-commerce sites - are you looking to improve your product search? It's time to look into AI. It's simpler than you think! In my new post I'll explain how to adopt AI incrementally, improve search relevance, customer experience, and conversion rate. arcturus-labs.com/blog/2026/01...
arcturus-labs.com
Incremental AI Adoption for E-commerce
Transform your e-commerce search from basic keyword matching to conversational AI—one step at a time. Learn how to incrementally adopt AI without overhauling your existing infrastructure, starting wit...
030
John Berryman @jnbrymn.bsky.social · 27/10/2025
𝐓𝐡𝐢𝐧𝐤 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐋𝐋𝐌 𝐚𝐬 𝐚𝐧 𝐢𝐧𝐭𝐞𝐫𝐧 𝐨𝐧 𝐭𝐡𝐞𝐢𝐫 𝐟𝐢𝐫𝐬𝐭 𝐝𝐚𝐲. The AI intern is: - 𝐁𝐫𝐢𝐠𝐡𝐭 𝐛𝐮𝐭 𝐧𝐨𝐭 𝐩𝐬𝐲𝐜𝐡𝐢𝐜 - 𝐏𝐫𝐞𝐟𝐞𝐫𝐬 𝐟𝐚𝐦𝐢𝐥𝐢𝐚𝐫𝐢𝐭𝐲 - 𝐍𝐞𝐞𝐝𝐬 𝐭𝐢𝐦𝐞 𝐭𝐨 𝐭𝐡𝐢𝐧𝐤 - 𝐇𝐚𝐬 𝐚𝐭𝐭𝐞𝐧𝐭𝐢𝐨𝐧 𝐢𝐬𝐬𝐮𝐞𝐬 - 𝐂𝐚𝐧 𝐛𝐞 𝐚 𝐜𝐨𝐧𝐯𝐢𝐧𝐜𝐢𝐧𝐠 𝐥𝐢𝐚𝐫 - 𝐇𝐚𝐬 𝐮𝐧𝐮𝐬𝐮𝐚𝐥 𝐯𝐢𝐬𝐢𝐨𝐧 - 𝐂𝐚𝐧'𝐭 𝐟𝐨𝐫𝐦 𝐧𝐞𝐰 𝐦𝐞𝐦𝐨𝐫𝐢𝐞𝐬 Full blog post is here arcturus-labs.com/blog/2025/10...
arcturus-labs.com
Context Engineering Requires AI Empathy
Transform your AI applications by thinking like an AI intern on their first day. This post reveals how empathy-driven context engineering leads to better LLM performance, covering everything from hand...
140
John Berryman @jnbrymn.bsky.social · 24/10/2025
I had an absolute blast talking about the "7 Deadly Sins of AI Applications" with Hugo Bowne-Anderson, and I'm really pumped to see it in text form now! (But I do think that neither cartoon John or cartoon Hugo are as handsome as the real deal.) hugobowne.substack.com/p/patterns-a...
040
Reposted by John Berryman
Doug Turnbull @softwaredoug.bsky.social · 28/07/2025
@jnbrymn.bsky.social and I are going to have a controversial discussion where maybe we don't agree maven.com/p/3ad9c7/ai-...
maven.com
AI Chat Startups - Is the bubble about to pop?
Investor money chases chat-style interfaces for every domain. I believe this is wrongheaded: ChatGPT and pals will do your domain better as they address more and more long tail-use cases, and better i...
021
John Berryman @jnbrymn.bsky.social · 25/06/2025
In this talk, I'll explore the state-of-the-art in LLM evaluation, covering modern techniques like LLM-as-Judge. I’ll also discuss the essential processes and culture shifts that are required to launch reliable AI products and drive ongoing improvement.
020
John Berryman @jnbrymn.bsky.social · 25/06/2025
Generative AI has made it easier than ever for companies to build products quickly. However, LLMs are inherently nondeterministic and unpredictable. Integrating LLMs into products demands an unprecedented level of quality assurance – requiring new strategies for continuous evaluation.
110
John Berryman @jnbrymn.bsky.social · 25/06/2025
I'm super excited to present to 7CTOs tomorrow on "Wresting with AI Evaluations" Register here: lu.ma/u9cp3exb
lu.ma
Cutting Through AI Hype with John Berryman (Part 2 of 2) · Zoom · Luma
Join us for the second in a two-part series that goes beyond AI buzz and into practical insight. This event will feature a dynamic blend of presentation, group…
130
John Berryman @jnbrymn.bsky.social · 12/06/2025
Then you can make a classifier that is tuned to whatever threshold matches a training set most accurately. Easy peasy. Full post here: arcturus-labs.com/blog/2025/03...
arcturus-labs.com
Supercharging LLM Classifications with Logprobs
Turn your LLM into a precision instrument for classification – no fine-tuning required. This post shows how to go beyond simple
010
John Berryman @jnbrymn.bsky.social · 12/06/2025
The idea is really simple – basically you just make a classifier the returns a single token – good/bad or red/green/blue or 1/2/3/4 – But rather than looking at that token you look at the probability of the tokens.
110
John Berryman @jnbrymn.bsky.social · 12/06/2025
You probably knew you could turn an LLM into a classifier. But the basic approach returns a hard classification – yes or no; good or bad. In this post I'll show you a simple technique to make your LLM classifiers more nuanced – they tell you HOW good or HOW bad something is.
110
John Berryman @jnbrymn.bsky.social · 06/06/2025
With the staggering amount of AI paradigms – Workflows, RAG, Agents – it seem impossible to get started learning AI and building your own projects. But once you realize that all of the ideas are built upon the same simple foundations, you become free to let your ideas flow.
maven.com
The Hidden Simplicity of GenAI Systems
At first, GenAI seems neatly split into agents, RAG, and workflows, but up close, it’s messy. This talk clears the fog by uncovering the small set of principles behind all major patterns. Once you gra...
000
John Berryman @jnbrymn.bsky.social · 08/05/2025
This should be a really fun talk, Shawn Simister and I have been recalling to ourselves how it all worked. What we realized is that we were doing an early form of LLM-as-judge back in early 2023. maven.com/p/da8264/how...
maven.com
How Evals Made GitHub Copilot Happen
GitHub Copilot stands as one of the first commercially successful generative AI applications. To make the product work, the copilot team had to invent evaluation methodologies with no existing bluepri...
021
John Berryman @jnbrymn.bsky.social · 05/05/2025
I'm looking forward to seeing you live-code a vector database. Lucky for me my only job for the call is to throw tomatoes. :D
031
Reposted by John Berryman
Doug Turnbull @softwaredoug.bsky.social · 10/04/2025
MAJOR ANNOUNCEMENT SoftwareDoug LLC has accepted $2K in seed funding from Turnbull Family Grocery Fund to bring you search/AI/RAG consulting, coaching, and training
162
John Berryman @jnbrymn.bsky.social · 07/04/2025
Read the full deep-dive on why visual reasoning is the next frontier in AI: arcturus-labs.com/blog/2025/03... Follow for more.
arcturus-labs.com
Visual Reasoning is Coming Soon - Arcturus Labs
From silly cat costumes to world-changing innovations, OpenAI's latest release marks the beginning of something extraordinary. The fascinating world of visual reasoning is emerging, where AI models wi...
010
John Berryman @jnbrymn.bsky.social · 07/04/2025
Picture this: AI solving physics problems by visualizing objects in motion, or predicting social interactions by imagining body language and facial expressions. That's where we're headed.
110
John Berryman @jnbrymn.bsky.social · 07/04/2025
The key insight: Just like chain-of-thought reasoning transformed how AI thinks through problems with words, visual reasoning will let AI work through problems by creating and analyzing sequences of images.
100
John Berryman @jnbrymn.bsky.social · 07/04/2025
Today's AI can put a detective hat on your cat. Tomorrow's AI will help design your garden, rearrange your furniture, and solve complex spatial puzzles by actually visualizing different scenarios and their outcomes.
110
John Berryman @jnbrymn.bsky.social · 07/04/2025
OpenAI just revolutionized image generation, but that's just the beginning. The real game-changer? Visual reasoning is coming - where AI won't just manipulate images, but will actually think through problems using visual simulation.
130
John Berryman @jnbrymn.bsky.social · 11/02/2025
@merlehazard.bsky.social well hi! Any new economic ballads recently? Surely you've no lack of inspiration right now. There's got to be some sort of ballad related to tarrifs.
120
John Berryman @jnbrymn.bsky.social · 11/02/2025
Exciting times ahead! As we enter a new golden age of creation with LLMs, I dive deep into RAG, AI-assisted coding, and the future of dev tools in my latest podcast with @dmitrykan.bsky.social for Vector Podcast
062
John Berryman @jnbrymn.bsky.social · 21/01/2025
👋 Want to know how to build AI agency into your application? Do it incrementally. Vlog here: youtube.com/watch?v=f5qW... and original post in video description. 🤙
youtube.com
030
John Berryman @jnbrymn.bsky.social · 20/01/2025
Thanks for reading our book and posting about it. I'm glad you got something from it!
110
John Berryman @jnbrymn.bsky.social · 18/01/2025
Building Reliable AI Apps? Start by Firing Yourself. Don't dive straight into complex AI agents. Start manual, document processes, then systematically automate each task. Full breakdown + case study: arcturus-labs.com/blog/2025/01...
arcturus-labs.com
Fire Yourself First: The E-Myth Approach to Iteratively AI App Development - Arcturus Labs
Discover how to build reliable LLM applications by applying the E-Myth's systematic approach to automation. Learn why starting with human processes and incrementally automating tasks leads to more rob...
010
John Berryman @jnbrymn.bsky.social · 27/12/2024
A perfect reply. Thank you Evgeny.
010
John Berryman @jnbrymn.bsky.social · 24/12/2024
I think I'll try this once a year until someone bites :)
010
John Berryman @jnbrymn.bsky.social · 21/12/2024
I was hoping that someone would fall into my trap actually :P – look up Cunningham's Law
130
John Berryman @jnbrymn.bsky.social · 20/12/2024
Turnbull's law: "The best way to get the right answer on the Internet is not to ask a question; it's to post the wrong answer." I love it! :D
260
John Berryman @jnbrymn.bsky.social · 20/12/2024
amzn.to/4fMj2Ef
amzn.to
Prompt Engineering for LLMs: The Art and Science of Building Large Language Model–Based Applications
Prompt Engineering for LLMs: The Art and Science of Building Large Language Model–Based Applications [Berryman, John, Ziegler, Albert] on Amazon.com. *FREE* shipping on qualifying offers. Prompt Engineering for LLMs: The Art and Science of Building Large Language Model–Based Applications
010
John Berryman @jnbrymn.bsky.social · 20/12/2024
Home for the holidays means cozy blankets, the scent of pine, and a tech book in the stocking. Prompt Engineering for LLMs—the perfect gift for the developer who finds magic in code and creativity in AI.
160
John Berryman @jnbrymn.bsky.social · 20/12/2024
I recorded 2 podcasts this week all about LLM application development. My recordings will be out soon, but until then make sure to check out these great programs: How AI is Built w/ @nicolaygerold open.spotify.com/show/3hhSTyH... Vector Podcast w/ @DmitryKan open.spotify.com/show/13JO3vh...
open.spotify.com
How AI Is Built
Podcast · Nicolay Gerold · How AI is Built dives into the different building blocks necessary to develop AI applications: how they work, how you can get started, and how you can master them. Build on ...
021
John Berryman @jnbrymn.bsky.social · 17/12/2024
Oops! Apparently the links above aren't shared. Try this instead www.oreilly.com/online-learn... – as a bonus, there are several other books to listen to besides ours.
oreilly.com
AI Audio Summaries - O'Reilly Media
Get in on the AI-generated conversation
031
John Berryman @jnbrymn.bsky.social · 17/12/2024
Of course, if you want to learn more you should actually just buy our book! 😂 amzn.to/4fMj2Ef And please follow me here on Bluesky as I plan to regularly produce information about LLM application development.
amzn.to
Prompt Engineering for LLMs: The Art and Science of Building Large Language Model–Based Applications
Prompt Engineering for LLMs: The Art and Science of Building Large Language Model–Based Applications [Berryman, John, Ziegler, Albert] on Amazon.com. *FREE* shipping on qualifying offers. Prompt Engineering for LLMs: The Art and Science of Building Large Language Model–Based Applications
110
John Berryman @jnbrymn.bsky.social · 17/12/2024
The second summary is 11min. It's made using NotebookML. drive.google.com/file/d/1yqZN...
drive.google.com
100
John Berryman @jnbrymn.bsky.social · 17/12/2024
The first summary is 6min. It's made using a tool that is being developed internally by O'Reilly. drive.google.com/file/d/1QOmp...
drive.google.com
100
John Berryman @jnbrymn.bsky.social · 17/12/2024
This is super exciting: Prompt Engineering for LLMs, has been selected for an upcoming beta feature on the O'Reilly Learning platform—AI-powered audio summaries. They've created two versions (next posts) tell me what you think.
190
John Berryman @jnbrymn.bsky.social · 16/12/2024
Look out folks! I am Liu-certified to be an AI Consultant now! Kidding aside, this course really helped me organize how I think about consulting. If you're thinking about tech consulting (even outside of AI), you should consider joining the next cohort maven.com/s/course/3aa...
030