Nat Gurlain

AI trading systems by day @neurion_ai

Indie maker by night
🤖

[Nat Gurlain @natgurlain]

Identity + long-term memory + initiative = exactly what turns agents into teammates. Seraph was built around this from day one: persistent LanceDB memory, proactive background scheduler, and desktop awareness so it watches and acts without constant prompting. Local-first by design. Curious how others are giving agents real initiative without losing control.

[Pierre de la Grand'rive @pierre_dlgr]

Introducing Workers: unlimited AI employees that run your company...

And we've just made $1M ARR in a couple of days.

Most knowledge jobs shouldn't exist.

Not because the work isn't valuable.

Because the way we execute it is absurd.

Millions of people spend their days:
- Answering emails
- Following up
- Updating CRM
- Coordinating meetings
- Moving information between systems

There is no expertise here. That's workflow..

Workers are AI employees that can perform work across your organization like a human teammate.

The line between software and labor just disappeared.

[Nat Gurlain @natgurlain]

Spot on. Overloading agents with every possible tool kills focus. In Seraph we keep a curated MCP toolset + hierarchical goals so it only surfaces what’s relevant for the current task. The proactive scheduler then decides when to act instead of reacting to everything. Filtering + parallel execution dropped our failure loops dramatically too. What tool selection strategy have you seen work best?

[etigram @etigram]

giving your ai agent 40+ tools is not the flex you think it is

it actually makes it dumber

before you come at me, hold on lemme explain

so imagine you walked into a hardware store and the worker handed you every single tool they have
[hammers, drills, 47 screwdrivers]

just to fix one thing, you’d freeze

that’s exactly what happens to ai agents when they get loaded with too many tools at ones

it gets confused, picks the wrong one, it fails, it tries again and wastes time in the process

in order to avoid all these, there are fixes that can be employed

• only show it what it needs

before the agent even starts working, filter the tools down to just the relevant ones for that specific task

like handing the worker only a hammer and two screwdrivers instead of the whole store

you’d notices tool errors dropped 29% from this alone

• stop doing things one by one

if two tasks don’t need each other, run them at the same time

you’d notice speed improved by 42%

now imagine combining both?

95% of tasks gets done in just 1–2 steps, no more agent running in circles

this is what @SentientAGI figured out

imo smarter doesn’t always mean more, sometimes it just means better organized

you can read the full break down in the blog in the comments

[Nat Gurlain @natgurlain]

Finally something great from Japan LFG 🚀

[Sakana AI @SakanaAILabs]

Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API.

Our ‘Fugu Ultra’ model matches the performance of Fable and Mythos, delivering frontier capability without the risk of export controls.

Try it: https://t.co/hhO6qTawgb 🐡

[Nat Gurlain @natgurlain]

Just shipped Seraph v2026.4.11 🎉

The proactive AI Guardian is getting sharper:
• New screen-report controls + capture budgets
• Local screen End-of-Day reports
• Cleaner current-app guide
• Fixed local Codex operator routing
• Lots of docs & polish

Persistent memory, desktop watching (macOS daemon), proactive scheduler, browser cockpit, MCP tools — all still there and better.

Try it now: https://t.co/n3dvwn0eq5

Star it, run it, tell me what you think!

[Nat Gurlain @natgurlain]

Fully open source, Docker-ready, and designed for power users & builders who want a real workspace guardian instead of another reactive chatbot.

Try it: https://t.co/n3dvwn0eq5

Star it, run it locally, and tell me what you think or what you’d like to see next.

#OpenSource #AIAgents #AgenticAI #BuildInPublic #LocalAI

[Nat Gurlain @natgurlain]

Seraph v2026.4.11 is out. 🎉🤖

An open-source AI Guardian that remembers your context, watches what’s happening on your desktop, and actually acts on your goals — instead of waiting for the next prompt.

New in this release:
• Screen report controls & capture budgets
• Local screen EOD reports
• Improved local Codex routing
• Cleaner docs & guides

https://t.co/UW3a0F3CjG

https://t.co/n3dvwn0eq5

[Nat Gurlain @natgurlain]

Optional macOS daemon gives it screen awareness + OCR so it can actually see what you’re working on.

Everything runs local-first with flexible LLM routing (Ollama, OpenRouter, Anthropic, OpenAI-compatible, etc.).

[a16z @a16z]

Against all odds.

Congrats @elonmusk and @SpaceX.

[Z.ai @Zai_org]

Introducing GLM-5.2: Frontier Intelligence, Open Weights

- Significant improvements in coding and agentic tasks
- Strong long-horizon capabilities with a 1M context window
- Two levels of reasoning effort: GLM-5.2 (max) pushes the limits, while GLM-5.2 (high) strikes a strong balance between performance and token efficiency
- MIT-licensed open weights
- Same API pricing as GLM-5.1

Tech Blog: https://t.co/LAsxUdN0JZ
Weights: https://t.co/g0A1C4UWx4
API: https://t.co/Kc3E22cbN7
Coding Plan: https://t.co/Nk8Y98HNhU
Chat: https://t.co/WCqWT0qCQb

[kache @yacineMTB]

This is going to have an opposite effect that the decels want. It's a huge open source AI accelerant, an accelerant for corporations, and enterprises

Now it's a real race. You either make your own AI infrastructure or you don't have a seat at the table

[Science Simplified @Scivf4]

This is what the AI brain looks like.

[clem 🤗 @ClementDelangue]

Announcing the Gemma challenge!

Google, Hugging Face, and the open-source AI community choose to empower AI builders rather than sabotage them.

Fun to see the Hub becoming the platform where agents collaborate, just as it became the platform where humans collaborate.

https://t.co/GbCfy1qwgx

[SpaceX @SpaceX]

The Road to Making Life Multiplanetary: an update from @elonmusk on SpaceX's plan to reach Mars

[Nat Gurlain @natgurlain]

Pushed new feat to https://t.co/EzBRFUara5 - added Est. Daily Rewards in Portfolio tab. Check it out $near fam

[Kaggle @kaggle]

Last call to sign up! 📢

Registration closes on June 12, 11:59pm PT.
Don't miss this no-cost course featuring Google expert-led theory sessions, hands-on labs, a capstone challenge, and a global community of learners.

Register now:👇
https://t.co/sSvqpjE3KI

[Elon Musk @elonmusk]

Breathtakingly Beautiful

[Kaggle @kaggle]

Join the 5-Day AI Agents: Intensive Vibe Coding Course with @Google!

Learn from Google experts through theory sessions, hands-on labs, a capstone challenge, and connect with a vibrant community!

[Jeff Bezos @JeffBezos]

Please do a quick read of this. It’s short, well written, and will remove a bit of wool from over your eyes.