Co-Founder @specstoryai. Organize D.C. AI tinkerers. Use https://xitter.gq/t.co/kg5JldHaiU, a calm agent multiplexer with familiar IDE features, for macOS.
Washington D.C.
Joined October 2015
- Tweets1.9K
- Following470
- Followers1.1K
- Likes1.4K
Greg Ceccarelli retweeted
Since developers aren’t coding anymore, their new job is to rotate secrets every single fucking day
Greg Ceccarelli retweeted
>junior dev at a mid startup
>wednesday morning
>just want to push a minor bugfix to our php app today
>open laptop, check slack
>company group chat is an absolute warzone
>CTO typing in all caps about php2shell
>fucking php
>whatever, server hasn't blown up yet, I just need to get my code in
>open gitlab to merge the PR
>forced emergency logout because of a critical gitlab security patch release
>fucking gitlab
>pipeline fails because of some weird regex compiler bug
>spend 40 minutes fighting CI/CD
>i can just use AI to help debug
>fire up opencode
>api keys nuked because they were breached
>fucking opencode
>i’ll just get fresh LLM api keys from our corp proxy, LiteLLM
>it won’t connect? It’s down?
>look up LiteLLM news: supply chain compromise
>fucking litellm
>cold sweat forming
>slack alert: db connection error
>check our database connection
>we host on TiDB Cloud
>they just announced a security breach, unauthorized access via stolen service account credentials, data copied
>fucking tidb cloud
>full panic attack incoming
>figure fuck it, just ship the container to Railway and go home
>open Railway dashboard
>bright red banner: "Security Disclosure: Rotate all environment variables immediately. All builds paused."
>fucking railway
>try to open the app URL in browser to see if anything is even working
>NET::ERR_CERT_AUTHORITY_INVALID
>eMudhra intermediate certificates compromised
>4
>instantly untrusted by every major browser
>fucking emudhra
>stare blankly at mac screen for 10min
>random pop-up
>restarting in 10min to fix a critical vulnerability
>fucking macos
>close laptop
>leave office
>never come back
Greg Ceccarelli retweeted
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.
So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
Judging by my tl there is a growing gap in understanding of AI capability.
The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.
But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.
So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.
TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
Greg Ceccarelli retweeted
Claude Code just shipped mods...
I made one that connects you to a multiplayer Doom server while Claude is busy.
Every other player in the match is a person waiting for their Claude to finish.
Greg Ceccarelli retweeted
It's been a while since we tweeted, but we've been shipping.
SpecStory CLI v2 → v2.13:
• 6 more agents supported
• Windows + WSL
• Secrets hidden by default
• Search & resume across computers
• Turn sessions into skills
Get the CLI:
curl -fsSL specstory.com/install.sh | sh
or github.com/specstoryai/getsp…
Greg Ceccarelli retweeted
I think AI coding is creating a new class of software developer.
People who’d never call themselves programmers, but can describe a system properly, test it, argue with the model, spot when something smells wrong and keep pushing until they get what they want.
They don't know every language feature or framework convention.
But they ship real software anyway.
Seems to be annoying the hell out of quite a few pre-ai programmers😂
Greg Ceccarelli retweeted
I've clearly found my new daily agent harness... or maybe agent harness harness.
If you haven't already, check out tortie.sh/
It's like a slim, agent driven IDE that has only what I realy need and no other cruft.
I can use whatever agent harness I want, it keeps everything super well organized, when anything happens to the Tortie process, the sessions don't die, I just start Tortie back and there they are ready to go. If the session itself has died, it's easy to restart.
But BEST OF ALL it's the only IDE I know about that has *phrase-based* red-green diffs. If you are using your agent to edit prose text, this is a game changer.
Pretend you're editing a proposal document. You say "fix the grammar and spelling errors".
If you're using Cursor or Zed then you get a Christmas tree – some text in every paragraph was changed, so you get 5 red blocks of text each followed by corresponding green blocks of text (left image). They at least highlight the changed words, but if the changes are complex then they are still basically incomparable.
But, Tortie gives you tight, inline diffs (right image). What's more, they aren't based upon git status, but based on whenever you open the file (which is more flexible IMO). And you can easily ⌥↑ or ⌥↓ to traverse through the edit sites and accept or reject them (you can also accept all).
This makes it really feel like I'm working with the agent when I'm handling prose text - I can see everything!
Greg Ceccarelli retweeted
an 18 year old kid built an imessage agent running a headless codex instance with astra where you sign in with your chatgpt account.
it works incredibly well. it can use apps too, not just a browser.
the primitives are getting absurdly powerful. sign up wasn't too bad either. i didn't need to set everything up again.
Instinct is painfully slow, so we built the world's best iMessage agent.
Introducing Companion
> gpt-6 astra by default
> bring your ChatGPT account (inherits context)
> 1000+ plugins from marketplace
> isolated computer with MCPs + CLIs
companion.result.dev 💬
Readers added context they thought people might want to know
OpenAI staff have stated this is not an approved use of Sign in with ChatGPT. OpenAI supports pure OSS clients and partnership products; others must contact them.
x.com/thsottiaux/sta…
help.openai.com/en/articles/20…
Greg Ceccarelli retweeted
There’s a phenomenon where talented individuals can achieve 10x their potential thanks to AI when working on their own thing
But when the same individual is put into a large organization, they at most increase their potential by 20% (and sometimes it’s even decreased)
This is why I’m seeing more and more talented people leave large companies. (The only exceptions are probably top AI labs like OpenAI/Anthropic)
Greg Ceccarelli retweeted
Terminals are too low fidelity, they say, so the web is a better platform for developer tools. I get the argument. But why then do I still prefer terminals for 80% of the work?
I think it's the constraints:
Terminals...
- have very limited space. You have to be diligent in showing *just* the essential info.
- are limited to (more or less) one-size plain monospaced text with a small color palette. Same benefit as above. No surprises.
- are keyboard-driven *first*. Modern terms and tools support mouse as well but usually as a bonus, not a requirement.
- are usually highly configurable and personalized. I pick the font, the colors, configure the tools, configure the hotkeys, etc.
- are power tools the same way a chainsaw is a power tool. Use it correctly, you can cut down any tree. Use it incorrectly, you can cut down your legs.
- are close to the ground. All the benefits of the "Unix philosophy" ring true still. Small, composable tools that do one thing and one thing well can pipe into and from each other, write to files, read from them, install packages, so on. Most tools weight stability over usability – but for an agent that has bash in its core training, that IS usability.
Greg Ceccarelli retweeted
I'm so freakin sick of how Claude talks.
It's not just annoying, it's infuriating.
The honest evaluation: it's the precise mechanism driving my current headache.
It's the load-bearing root cause of my daily stress.
It's genuinely painful, and that matters.
Greg Ceccarelli retweeted
I wanna be a HARDCORE AGENTIC ENGINEER.
Pumped to be taking Hardcore Agentic Engineering for Builders Who Ship, @gregce10's three-week course.
Greg is a beast. He led data science at GitHub, went on to become Chief Product Officer at Pluralsight, and is now co-founder and CPO of SpecStory. When he joined us for Show Us Your Agent Skills, he showed how he mined 516 of his own agent sessions for reusable practices and built a harness that keeps agents working until independent checks confirm the work is done.
That’s the level this course is operating at, and that’s why I’m taking it.
Greg is offering my network 35% off with the code HUGO if you want to join next week.
This is absolutely not a course for beginners. It’s for people already using Claude Code, Codex, and other coding agents seriously who want to go much deeper.
check it out: maven.com/s/course/03fa3c209…
Oh, also this how Greg and I look when we're vibe coding an app to view the JFK files:
Greg Ceccarelli retweeted
I’m significantly older than you. I started coding in the late 60s. My current strategy is to not read any of the code written by my agents. That’s the only way I can take advantage of their productivity. What I do instead is to surround the agents with extreme constraints. Unit tests, gherkin tests, QA procedures, quality metrics, mutation testing, test coverage, and a plethora of others. In the end, I have very high confidence in the code they produce because they’ve had to run the gauntlet of all of my constraints and tests.
Greg Ceccarelli retweeted
Thanks to Greg and @HanchungLee for such a wonderful skills-sharing session.
Watch the full episode:
youtube.com/live/kfCi2EBu-nc
