Garrett Tacoronte

I grow companies through AI, people, technology, sales, and marketing

[Garrett Tacoronte @10exdev]

I blew through my weekly budget in the last 36 hours with claude max 20x, so you can say I got some practical use out of it. The new Fable 5 model is really good! Great job Anthropic
But strangely, it feels like Fable is on par with last week's opus 4.8, if not just slightly better. Today, opus 4.8 feels nurfed for sure....
I wonder if these models are only barely noticeable at being better, so they just nurf the most recent previous model to make the gap seem bigger than it is...

[Garrett Tacoronte @10exdev]

@josevalim I had to build out a skill to specifically prevent it from doing this
And it still does it lol

[Garrett Tacoronte @10exdev]

Sort by the columns. Use the bottom 25 percent for coaching your AMs. Now you don't need to listen to every call, just the flagged ones. Takes under five minutes to run the first time and so worth it!

[Garrett Tacoronte @10exdev]

Your account managers handle 40 client touchpoints a week. No one can listen to enough of them to know what calls were good or bad, or which AM is good or bad, or which clients are at risk of leaving you.

[Garrett Tacoronte @10exdev]

Export the last 30 AI summaries from your call tool. Paste them into a sheet and add columns for talk ratio, objection mentions, and whether a next step was booked on the call.

[Garrett Tacoronte @10exdev]

@davydog187 @wojtekmach @elixirlang Amazing work! This looks fantastic and the landing page looks really nice too 🔥

[Garrett Tacoronte @10exdev]

@josevalim Yup, this would be a great addition!

[DHH @dhh]

Mise makes dev life so much simpler. A single env manager for Ruby, JavaScript, Go, and all the modern AI tooling. Every project can have their own versions. Stable system packages can be separated from high-churn AI tooling. Thrilled to sponsor @jdx in this mission!

[Garrett Tacoronte @10exdev]

😂

[Justine Moore @venturetwins]

Me using Claude Opus 4.8 to rename a file

[Danila Poyarkov @dan_note]

A lot of people expressed interest, so I decided to open source this earlier than planned.

Vibe combines the vibe-coding tools I’ve been building into an Elixir-focused coding agent. But it is not only about coding: the goal is to support background tasks, long-running agent workflows, and access
through Telegram and other gateways.

Some highlights:

— Elixir eval is the primary tool interface. This makes it possible to unify many actions behind one composable interface, in a Unix-like way, without constant serialization churn.

— Sessions are stateful. Similar to Livebook, the agent can store and reference intermediate results naturally.

— Unlike most agent harnesses, Vibe is built on OTP, so it can deeply introspect and supervise its own internal state.

— Agents can launch other agents. Subagents are supervised OTP processes with their own sessions, which enables more sophisticated long-running and parallel workflows.

— Vibe supports remote access over SSH and Erlang distribution, so agents can connect to remote nodes and communicate across machines.

— It has a TUI, a LiveView web interface, persistent sessions, SQLite-backed storage, semantic events, plugins, skills, subagents, gateways, and local telemetry.

— The architecture is meant to be hackable by the agent itself: it can inspect, patch, verify, and hot-reload parts of Vibe, while tools like Reach help keep boundaries clean.

This is still an early research preview. Use it at your own risk, expect rough edges, and be careful.

https://t.co/2DqgVdx6yo

[Garrett Tacoronte @10exdev]

@GergelyOrosz Apple Studio Display was a massive upgrade that I don’t think I can ever go back from.
I have a hard time working on my laptop anymore because it’s not as nice.
It 100% was worth it for me and I highly recommend it

[Garrett Tacoronte @10exdev]

Daily AI summaries change the game. One inbox scan shows tone shifts, shorter calls, and competitor mentions before they turn into lost revenue.

[Garrett Tacoronte @10exdev]

Most agencies treat client calls as one-off events. The account manager hangs up, notes the action items, and moves on.

[Garrett Tacoronte @10exdev]

Six weeks later the same client cancels. The warning signs were in those calls the whole time. They just never got reviewed as a pattern.

[Garrett Tacoronte @10exdev]

Six months later the same attribution gaps reappear because that hire treated the request as a task instead of fixing how the data gets created in the first place.

[Garrett Tacoronte @10exdev]

Agency owners hire the technical person who ships the new dashboard in two weeks.

[Garrett Tacoronte @10exdev]

Would you rather keep explaining the numbers or let the output speak for itself on the next call?

[Garrett Tacoronte @10exdev]

Your account managers rebuild the same client ROI summary from scratch every month because the data lives in four places.

[Garrett Tacoronte @10exdev]

Run it on a single client first. Takes under five minutes. This is going to be the biggest ROI indicator to your client showing real performance from your work.