Greg Ceccarelli

Co-Founder @specstoryai. Organize D.C. AI tinkerers.

Use a calm agent multiplexer with familiar IDE features, for macOS.

[Greg Ceccarelli @gregce10]

hAvE yOu hEaRD aBoUt: omniagent, paseo, orca, t3code, herdr, cmux, etc etc?

ofc but I drive my with agents sessions with a cat.

👉https://t.co/Xl5Qd3PgN4

a calm agent multiplexer that feels like vscode, doesn't make you learn tmux and does everything it can to save your work

[Greg Ceccarelli @gregce10]

I wrote about "goal engineering" back in May but I actually have significantly different thoughts now and think its irrelevant.

said differently, the process outlined below is dated and not required if you use structured workflows with the latest models (i.e. Opus 5.5)

https://t.co/cqk6TfEflZ

[Greg Ceccarelli @gregce10]

Field notes with worked examples on a goal engineering workflow I've been running and am excited about: instead of writing prompts & specs, I now focus on two checked-in markdown artifact per round of work.

The "goal" is capped at 4,000 chars (the same limit Codex's /goal command enforces). The "rider" is unbounded, with +/- eleven phases and named depth tests. I write them via a Skill included in the article. Looking for long running agentic turns, this is for you.

https://t.co/avLrfgZVjb

[Greg Ceccarelli @gregce10]

millennial kryptonite is having an out of town boomer parent's online account be compromised and being physically and emotionally incapable of helping them

[Greg Ceccarelli @gregce10]

@bossriceshark On one hand I agree and on the other, it agrees too

[Greg Ceccarelli @gregce10]

Word to the wise: this is NOT a full account reset.

Its just for Fable usage.

Don't be like me (without reading closely) and bash reset. I did this: dropped from 100% consumed to 92% because honestly who was using Fable 5.1 up until today when Opus 5.5 has been so grand to consume their usage.

So really not that generous, definitely misleading and yet another way that Ant sucks at comms.

[Boris Cherny @bcherny]

We wanted to do more for the people who’ve been with us the longest. So with Fable 5.2, the longer you’ve been subscribed, the more Fable resets you get.

[Greg Ceccarelli @gregce10]

@bossriceshark So perplexing and honestly... like wtf. When have their been MODEL specific usage resets? New class of reset I guess but srsly.

[Greg Ceccarelli @gregce10]

@intellectronica @hugobowne https://t.co/9f7mi2CXe0

[Thorsten Ball @thorstenball]

Yup, we have SOC2 and push to main:

https://t.co/Iws7OumCz7

[Greg Ceccarelli @gregce10]

@hugobowne many are moving this way :)

[Greg Ceccarelli @gregce10]

@verrsane Should pay for Cleanshot X. It’s so good

[Greg Ceccarelli @gregce10]

Moments ago there was a deafening roar outside… Airforce 1 overhead my roof deck by Catholic U in DC.

[Greg Ceccarelli @gregce10]

Highly recommend you start doing this. Here is an example of how long an org and billing spike took in precise "agent hours" (new estimation unit to replace eng weeks?)

Curious about how to get started? Read: https://t.co/L2fHnpb8KQ

[Greg Ceccarelli @gregce10]

Using claude code workflows extensively with access to session traces and message timestamps has an unexpected fringe benefit: am able to produce very precise estimates of total wall clock time required to spec, build and fully verify features.

Before: claude was like "Oh this plan will take 2 to 3 engineering weeks" and we all knew that was wrong.

Now: claude is like it will take 7 hours of agent build and integrate implementation time, 5 - 6 hours to verify, fix and reverify and 45 minutes to release to prod via your process based on the last 60 phases you ran and my calculation of spec complexity, files touch, verification scenarios to run through for this change and your CI/CD timings.

Its been accurate within +/- 10 pp over the last few weeks.

[Greg Ceccarelli @gregce10]

Also this

[Watcher.Guru @WatcherGuru]

JUST IN: Google $GOOGL crashes over 13% at open.

[Greg Ceccarelli @gregce10]

@BEBischof oh yea human, you want a combobox for your filter option, huh? how about we use this mildly infuriating downward facing chevron with just too little right padding. I bet you'll like that!

its claude's version of civil disobedience

[Greg Ceccarelli @gregce10]

@Zimmerman895 Sub agents read them, different models are in use everywhere that aren’t so bitchy vigilant, people paste them directly into prompts all the time

[Greg Ceccarelli @gregce10]

You would not even believe the amount of credentials / keys (aws, gcp, azure, saas vendors, oai, ant, etc) exposed to models via agents on a daily basis in agent session logs.

I don't even think you can even fathom the volume of exposure.

[Greg Ceccarelli @gregce10]

@kartikb753 All of the above + people just copy pasting

[Greg Ceccarelli @gregce10]

@joehelle Gus point

[Greg Ceccarelli @gregce10]

@mstockton see my bio ;)

[Greg Ceccarelli @gregce10]

@UK_Daniel_Card well just like a said…! ;) what would you like to know?

[Greg Ceccarelli @gregce10]

@_rybaz The intelligent choice!

[Greg Ceccarelli @gregce10]

@JonCrosby_ Think about the massive number of New people who start using these products everyday.

I can bet they will keep pasting keys in text boxes!