Jag

Founder/Builder/Educator
I’m not a financial advisor.

[Jag @jsb81]

@patloeber Superb. Gemini 3.6 flash is very fast and great working with large documents and pdf files.

[Google Cloud Tech @GoogleCloudTech]

This summer, shift into GEAR—the Gemini Enterprise Agent Ready program! Follow along for 12 training and learning opportunities to build agentic AI into your workflow ↓

[Supabase @supabase]

📦 The new official Server Package makes Auth easy in Edge Functions!

It automatically handles:

✅ JWT Verification
🔑 Authenticated queries
👮‍♂️ Admin client
🌐 CORS

And is compatible across Deno, Cloudflare Workers, Vercel Functions, Hono and Bun!

[Google Cloud Tech @GoogleCloudTech]

Build search and reasoning capabilities for your apps with Gemini Enterprise Agent Platform.

Take this course to learn how to do just that—and work with Flutter, Google's UI framework used to build apps for multiple platforms from a single codebase → https://t.co/Otg5KxmA8Q

[Ritwik Pavan @ritwikpavan]

NEW: @thelightphone just launched a $299 flip phone for people tired of doomscrolling.

Light Flip is a modern minimalist phone with 5G, a physical keypad, and no touchscreen, browser, email, social feeds, or ads.

-2.8-inch OLED display on the inside
-Physical T9 keypad with predictive texting
-No touchscreen and no external screen
-Notification light shows alerts while closed
-Runs LightOS with calls, texts, calendar, maps, music, podcasts, and basic tools
-5G, eSIM, nano SIM, USB-C, Bluetooth, and headphone jack
-Rear camera, replaceable battery, and six color options

Priced at $299. Ships April 2027.

[Supabase @supabase]

Introducing Supabase Pipelines.

Your app runs on Postgres, but your analytics usually live somewhere else, like BigQuery. Pipelines keeps that data in sync for you in near real time, with no pipeline to build or maintain.

[Burke Holland @burkeholland]

Prompt caching in @GitHub Copilot in 13 minutes.

This video almost did me in. Easily the hardest topic I've ever tried to cover.

[Jag @jsb81]

Why are businesses so focused on token spend when open-source models can be deployed on-premises or in the cloud without provider-based per-token pricing?

Wouldn’t it be more meaningful to measure the overall ROI across compute infrastructure, data governance, AI governance, agent development, maintenance, and the business value generated?

Interested to hear how others are approaching this.

[Linear @linear]

Introducing Loops: Recurring workflows that Linear Agent can run for your team.

[Cursor @cursor_ai]

We had a team of agents rebuild SQLite from its 835-page manual.

It created a replica in Rust which passed 100% of a held-out test suite.

Interestingly, cost varied 15x depending on which model mix we used.

[NetworkChuck @NetworkChuck]

Worried about AI taking your job? You need a Telos File. 📂

It’s a single Markdown document with everything about you. Then, we run it through a personal pentest; challenging your goals, values, identity, blindspots, and vulnerabilities. In this video, I’ll show you how to use it: https://t.co/k7er3BUtwY

[Reads with Ravi @readswithravi]

Clarity comes from action ‼️

[Jag @jsb81]

@nutlope looking forward to repo. Thanks Hassan.

[Hassan @nutlope]

Hacking on an interactive 3d map of AI startups in NYC!

Launching soon. Will go over how I built it & will open source it so people can add their own startups.

[Ryan Hart @thisdudelikesAI]

Microsoft has released an open-source tool that helps teams learn ontology design before choosing a knowledge graph platform.

It is called Ontology Playground.

The project is a fully static React app, which means it does not need a backend, account system, database, or hosted service to run.

The goal is simple:

Help people understand what goes inside a graph before they buy or build the graph database.

Ontology Playground includes six pre-built domain ontologies:

→ Retail
→ Healthcare
→ Finance
→ Manufacturing
→ E-Commerce
→ Education

Each one gives users a starting point for understanding entities, relationships, properties, and how domain knowledge gets structured.

The app also includes a live visual designer, structured learning paths, hands-on labs, and RDF/XML export for Fabric IQ.

Because it is static, it can be deployed almost anywhere.

No backend.
No vendor lock-in.
No platform commitment upfront.

This matters because many teams jump into knowledge graphs too early.

They focus on the database first.

But the harder question is usually:

What should the graph actually know?

Ontology Playground teaches that layer first.

[Jag @jsb81]

Interesting perspective. It looks like the $600,000 is only the savings on software license fees. The actual cost to build and support the system would be a separate expense. Keeping these systems running—which usually involves specialized Salesforce partners—is rarely cheap and depends a lot on the partner's resources. A common issue with enterprise SaaS is that companies are recommended to change their business processes to fit the software. This can make it hard for the team to adopt it. Building a custom solution fixes this problem. In the AI era, managing a custom build does not have to be a challenge if proper data and AI governance is implemented from the start. With good governance and structured engineering, those maintenance costs can be managed well. For teams looking to do this safely, Microsoft has a helpful framework on how to Build an agentic Center of Excellence that gives a great roadmap for managing these projects.
https://t.co/zZuX6ClLK8

#softwarefactory #googlecloud #cloudflare @GoogleAIStudio

[Harry Stebbings @HarryStebbings]

"We replaced Salesforce with a vibe-coded CRM built for our own workflows.

The custom system integrated our AI agents more effectively, worked better for the team, and made Salesforce unnecessary.

That decision cut a $600,000 annual software bill to zero."

Is this an anomaly or the start of a much larger trend @chamath @Avishai_ab @jasonlk @benioff

[beamnxw ./ @beamnxw]

A Google engineer showed how to fine-tune a tiny LLM directly on a phone and raise its accuracy from 46% to 90% in just 21 minutes

Here is the approach:

1\ Choose Gemma 270M
2\ Generate synthetic data for your specific task
3\ Fine-tune the model using LoRA
4\ Quantize it to int4
5\ Deploy on a Pixel phone and achieve up to 2000 tokens per second

The full stack = Gemma 270M + synthetic data + LoRA + int4 quantization + on-device runtime

This lets you run a compact, high-performance AI agent offline right in your pocket

Watch, save, and try training your own compact AI agent today

[Shruti @heyshrutimishra]

Sundar Pichai just reminded everyone that Google was built on open source.

He personally worked on Chromium, Android, and Kubernetes before becoming CEO. Three systems that power billions of devices today.

Now Google is applying the same philosophy to AI. Their Gemma models are updated year after year, designed to run on edge devices rather than requiring massive cloud infrastructure.

The question he got was why not release a large open source frontier model. His answer was revealing: frontier training requires enormous capex investment, and Google is putting billions into R&D to stay at the frontier.

The open source AI landscape is quietly consolidating around a few players who can sustain the investment. Google's two decades of open source credibility gives them a position that's hard to replicate.