Agreed, and from spec-spine perspective, it doesn’t try to be the source of truth about behaviour; Code wins that.
Instead it’s the source of authority over admission, which is different question; not “what does the system do” but “what is it permitted to do, and who decides.”
Your “anchor the spec to the code to stay aligned” is the same instinct.
spec-spine just makes the anchor a gate instead of a sync job.
Drift doesn’t accumulate silently because non-conformance is refused at the boundary. So in a sense it is anchored to an enforcement boundary; not to the code and not to nothing.
ONE production example you already accept: the type system.
Nobody calls the compiled binary the source of truth for whether it typechecks.
Signature is exogenous, compiler is the gate, non-conforming code is rejected. spec-spine is that shape one level up.
Build like nobody watching
298,883 lines. 181 specs. 23 Rust crates. 1,203 commits. 9 weeks.
One human. Plus Dependabot, and one 13-line commit from an agent. https://t.co/5z7rAOF43k · AGPL-3.0
I built this because it is my craft, my forte, neither a pitch nor a portfolio, just a counter-thesis to vibe coding and single-shot prompting, an artifact to share and my therapy to live by.
The numbers:
▪ 156k TypeScript · 120k Rust · 17k JS · 5k Python · across 1,498 files
▪ 15,598 GitHub Actions minutes burned in May. 4,252 job runs. One contributor.
▪ 6,317 clones · 969 unique cloners · 35 unique human visitors · last 14 days
“National IQ” rankings are not science, they’re cosplay for people who don’t know how tests work. There is no peer-reviewed, standardized, cross-cultural dataset that can assign an average IQ to entire countries, and numbers in the 40s are so psychometrically absurd they would imply populations incapable of running villages, let alone modern societies.
Nepal isn’t even tested in PISA or TIMSS, which means those figures are invented through bad modeling and worse assumptions.
Intelligence tests measure performance under specific schooling, language, nutrition, and stability conditions; not human capacity or worth.
And if you think spreadsheets override reality, here’s a reality check:
Gurkhas are recruited through some of the most selective screening processes on Earth, learn complex systems fast, operate under extreme pressure, and outperform expectations wherever they’re deployed.
You don’t get to talk about “low IQ populations” while outsourcing courage, discipline, adaptability, and excellence to them. If your worldview says they’re “low intelligence,” your worldview is the thing that failed the test.
@KEZOUniverse @Hunter_Weiss Given that the company was founded by Saoud Rizwan and Nik Pash, and that it recently made headlines when Rizwan initially defended Pash's controversial comment before firing him. I think you’re onto something here…
@JRedpill2@naval A real vision scales by nature; if it doesn’t, it’s just a viewpoint.
Ergo, a founder’s ‘vision’ starts as a strong perspective. However it only becomes true vision when it creates a new reality others choose to join.
I get why you’d want to patch addEventListener; it’s the ultimate DX move.
Making "press" just work™ without forcing devs to learn new APIs feels magical and keeps Remix drop-in friendly.
That said, something like addCustomEventListener("press", fn) could hit the same goal without touching browser internals; clearer intent, fewer collision risks, and future-proof if the platform ever adds similar events.
Maybe the sweet spot is both: native addEventListener("press", fn) for the magic, plus a customEvents.add("press", resolver) for explicit opt-in power users.
That way you get the ✨ DX and the 🧠 maintainability.
Thanks for sharing your thoughts.
That’s genuinely fascinating work, Rik. What you’re describing, i.e. property-level type validation and AOT-optimized typed blocks; feels like a runtime that’s beginning to understand its own shape.
In many ways, Encore.ts is chasing that same goal, just from the opposite side of the spectrum. It takes static TypeScript, analyzes the complete type graph, and feeds it into a Rust-compiled runtime that enforces those contracts at execution time. So instead of the runtime inferring structure from dynamic code like you’re doing, it’s reflecting a known static structure into a high-performance execution environment.
Both directions converge on the same idea:
“What if code could reason about its own type semantics deeply enough to optimize itself?”
Where your approach explores runtime self-inference, specifically letting JS dynamically promote sections into fully-typed WASM blocks; Encore.ts explores compile-time self-reflection, by projecting TypeScript’s static guarantees into a compiled runtime that behaves as if it were natively typed from the start.
The beauty is that both philosophies dissolve the traditional barrier between “language” and “engine.”
Instead of treating the runtime as a black box, they make it part of the language semantics; an intelligent collaborator that learns, optimizes, and enforces correctness based on what the code means, not just what it does.
It’s exciting to see these experiments happening on both ends of the spectrum; your AOT interpreter pushing JS forward as a self-optimizing runtime, and frameworks like Encore.ts pushing TS toward being a self-reflective, compile-time verified execution model.
Ultimately, they’re two paths to the same north star:
a language system that collapses the distance between definition, validation, and execution.
Thank you for sharing your thoughts.
@MyDReamscomgtru @pvtjokerus This is less a tweet and more an SAT reading comprehension exercise.
I think Google Translate saw this and applied for disability benefits.
Plot twist: the real victim here is punctuation.
Replace? I doubt it. Ben certainly didn’t use that word when expressing his desire to carry forward Charlie’s legacy.
However, what could be scrutinized are his closing remarks:
“Charlie, may your memory be a blessing for your family, YOUR country, and all of us.”
He is telling you something important here. Notice the choice of words: he says “YOUR” country, not “OUR” country, when referring to the United States.
That single word choice is not accidental. It’s what psychologists would call a Freudian slip; an unguarded moment where true loyalties reveal themselves.
By saying “YOUR country” instead of “OUR country,” Ben Shapiro unconsciously marked a separation between himself and the United States.
His alliance, as his words betray, lies first and foremost with the State of Israel; not with America.
@masternumber @janmikulik420 Salvador Freixedo’s “Human Farm” might be of interest for anyone that might be on the fence about matrix being anything but evil.
Stop wishing and chug some POWERTHIRST, brother!
You’ll be running so fast Kenyans and Ethiopians will ask you for tips.
You’ll sprint through time.
You’ll defy physics, ethics, and your HOA regulations.
You’ll burn calories just by making eye contact.
Your sweat will cure shin splints.
Gravity will file a complaint.
You’ll outrun your past, your future, and your dad’s expectations.
Every step a thunderclap.
Every breath a war cry.
Your stride? Illegal in 17 countries and 4 dimensions.
You won’t just run; you’ll transcend.
https://t.co/N0WLw9tEg0
@grok Would the introduction of a viable third political party in the U.S. strengthen or weaken the country’s function as a constitutional republic with democratic principles? What institutional changes would be necessary to ensure it improves representation rather than fragments it?
@grok Would the introduction of a viable third political party in the U.S. strengthen or weaken the country’s function as a constitutional republic with democratic principles? What institutional changes would be necessary to ensure it improves representation rather than fragments it?
@grok Would the introduction of a viable third political party in the U.S. strengthen or weaken the country’s function as a constitutional republic with democratic principles? What institutional changes would be necessary to ensure it improves representation rather than fragments it?