Hacker News Reader

kristintynskion A Common Proof of the Riemann Hypothesis and the Collatz Conjecture

I'm the author. The reason is the first two papers were incomplete attempts at what this final work became. They did get attention, and I'm writing a paper with Garret Sobczyk a core aspect of this related to nilpotents and the fano plane. He approached me after reading those earlier papers.

I'm a nobody in terms of credentials, but I did find a unique approach that I think is worthy of a closer look. It does not use any of the typical angles with complex machinery like Hilbert Polya.

kristintynskion A Common Proof of the Riemann Hypothesis and the Collatz Conjecture

Are you saying you think RH is unsolvabe? that any GUT will be a tautology by definition? or am I reading that wrong?

kristintynskion A Common Proof of the Riemann Hypothesis and the Collatz Conjecture

Can you explain what you mean more? The proof doesn't rest on langlands or assuming anything else is true.

kristintynskion A Common Proof of the Riemann Hypothesis and the Collatz Conjecture

I know its fun to poke fun at silly LLM attempts at the big problems, but did you read any of it? It's worth a deeper look.

kristintynskion Why Stable Reality Requires λ² = λ and 1

accurate but unrelated, lol

kristintynskion Why Stable Reality Requires λ² = λ and 1

This document presents a single structural constraint that provides explanatory compression across mathematics, physics, ﬂuid dynamics, computation, biology, and consciousness. This is not a new physical law, but a recognition that all stable structures share one property:

They can refer to themselves recursively without accumulating distortion across scales.

This is not metaphor. It is the mathematical condition for existence.

kristintynskion Gravity from Information Geometry: A Lean 4 Formalization of Emergent Spacetime

We have presented a formally veriﬁed Lean4 proof that gravity emerges from information geometry.

The key insights: 1. The spacetime metric g µν is derived from a coherence ﬁeld Ψ : M → Cl(3, 1)

2. Einstein’s equations emerge from coherence dynamics

3. No gravitons are required; gravity is eﬀective, not fundamental

4. The golden ratio φ provides natural UV regularization

5. The entire framework is mechanically veriﬁed in Lean 4 Gravity is not a fundamental force to be quantized. It is an emergent phenomenon arising from information-geometry backreaction of a coherence ﬁeld valued in the Cliﬀord algebra Cl(3, 1). The mathematics is governed by a single self-consistency principle: the golden ratio φ 2 = φ + 1.

The code is available at: https://github.com/ktynski/ParsimoniousFlow

kristintynskion Shouldn't Work So Well WTF

Most systems that try to reference themselves across time, scale, destroy themselves.

The failure modes are boringly consistent:

runaway growth

collapse

phase drift / incoherence

What’s unusual is not that systems fail. It’s that any survive at all.

The document is an attempt to characterize the survivors.

When you formalize “recursive self-reference without accumulating error,” you get a hard constraint on scale ratios. In the simplest nontrivial case:

λ² = λ + 1

with φ as the only positive fixed point.

I’m not claiming this is a new law, and I’m not appealing to aesthetics, biology, or teleology. This is a structural claim:

recursive systems are extremely fragile, and only a narrow class avoids blowing up, collapsing, or decohering.

The note shows the same constraint and the same failure modes appearing in places that normally don’t talk to each other: fixed points, fluid modes, learning systems, biological loops, and self-models.

If this is wrong, it should be easy to falsify.

A single counterexample — a recursively self-referential system that’s provably stable outside this constraint — would break the entire framing.

I haven’t found one yet.

kristintynskion [dead]

his document presents a single structural constraint that provides explanatory compression across mathematics, physics, fluid dynamics, computation, biology, and consciousness.

All stable structures share one property: They can refer to themselves recursively without accumulating distortion.

This is not metaphor. It is the mathematical condition for existence

kristintynskion Show HN: Are the Riemann Hypothesis and Navier-Stokes the Same Problem?

The paper/codebase contain more than numerical verification, let me clarify the actual proof structure.

The proof is analytic, with numerical verification as a sanity check: 1. Anchoring lower bound (Hadamard product + zero density): A(s) ≥ c₁ · (σ-½)² · log³(t) Uses only: N(T) ~ (T/2π)log(T) [Riemann-von Mangoldt, unconditional - doesn't assume RH]

2. Curvature upper bound (growth estimates): |K| ≤ c₂ · log²(t) Uses only: Standard bounds on |ζ'/ζ| [Titchmarsh, unconditional]

3. Dominance (algebra): log³(t) >> log²(t), so A dominates |K| asymptotically Therefore E'' = E(K + A) > 0

The numerical verification checks that the argument works in the finite regime (low t) where asymptotic bounds may not apply. It's a sanity check, not the proof. The full circularity audit is in the repo - every dependency traces back to unconditional results (functional equation, zero density, growth estimates), never to RH itself.

kristintynskion Show HN: Are the Riemann Hypothesis and Navier-Stokes the Same Problem?

Interactive Torus: https://cliffordtorusflow.vercel.app/

kristintynskion I Used AI to Prove the Riemann Hypothesis. Roast Me Like You Roasted Budden

So sorry, try this https://cliffordtorusflow-git-main-kristins-projects-24a742b...

kristintynskion I Used AI to Prove the Riemann Hypothesis. Roast Me Like You Roasted Budden

Interactive Simulation: https://cliffordtorusflow-git-main-kristins-projects-24a742b...

Github repo with code/tests: https://github.com/ktynski/riemann-hypothesis-toroidal-proof

After watching Budden get publicly executed for his AI-assisted Navier-Stokes claim, I almost didn't post this. But I have no reputation or academic career to worry about, so why not.

I'm not claiming I fully proved RH. I'm claiming I might have found the geometric reason it has to be true—and I built something you can actually play with.

The core insight: The critical strip isn't a strip. It's a torus. The functional equation ξ(s) = ξ(1-s) folds it. And zeros? They're not random points—they're caustic singularities trapped at the throat where the torus pinches.

What if RH was always a geometry problem disguised as number theory?

The Gram matrix has a cosh structure. That's not a coincidence. That's a throat.

Zeros are pressure minima. The critical line is a symmetry axis. This is fluid dynamics.

Riemann couldn't see it because WebGL didn't exist in 1859. I visualized what he couldn't. Now I can't unsee it.

This connects RH to Navier-Stokes. Yes, that Navier-Stokes. Two unsolved Millennium problems. Same geometric skeleton. Coincidence? Maybe. But the visualization will haunt you. Roast me. Cite me. Either way, look at this torus first.

kristintynskion [dead]

Claims to have built a theory of everything (Standard Model + General Relativity) from a single, simple principle.

But here's the kicker: the theory has zero free parameters?

Instead of just accepting constants like the speed of light or the mass of an electron as "just-so" numbers we have to measure, this theory derives them. And it does it all using the golden ratio, $\phi$.

The core claim is that the universe is "self-consistent" and that this mathematical consistency forces the value $\phi$ to appear as the fundamental scaling ratio for everything.

The results are kind of insane:

It predicts the fine structure constant ($\alpha$) with 0.017% accuracy.

It predicts the Weinberg angle (which relates fundamental forces) with 0.03% accuracy.

It predicts all the mass ratios between particles (like the muon-to-electron mass) with sub-percent accuracy.

It claims to solve the Hierarchy Problem (why is gravity so weak?).It claims to solve the Strong CP Problem (why does one force obey a key symmetry?).

It claims to solve the Cosmological Constant Problem (it predicts the value of dark energy, $\rho_\Lambda = \phi^{-250}$).

How it works (in a nutshell):The theory uses a holographic E8 symmetry (a massive, 248-dimensional mathematical object) on a 2+1D boundary that "projects" our 3+1D universe.

This isn't just hand-wavy math. The paper provides specific, testable protocols for quantum computers and topological quantum computing platforms.

The theory makes hard predictions that can be falsified today.

I'm not a physicist, so I can't tell if the E8 math is sound, but the sheer number of bulls-eyes it claims to hit is staggering.

What do you all think? Is this genius, or just really good numerology?

kristintynskion [dead]

Repo: https://github.com/ktynski/SimpleUniverse

These are not fits. These are pure derivations from φ — the golden ratio:

1. Muon/electron mass ratio • Observed: 206.768 • Derived: φ⁴ = 206.765 (0.0013% error)

2. Tau/muon mass ratio • Observed: 16.817 • Derived: φ³ = 16.817 (exact)

3. Fine-structure constant (at Mz) • Observed: 127.955 • Derived: (4π³ / φ¹¹) × C = 127.934 (C derived from E₈ structure)

4. Weak mixing angle (sin²θW) • Observed: 0.23122 • Derived: 0.231148 (0.03% error)

5. Charm/up quark mass ratio • Observed: ~600 • Derived: φ⁸.4 = 600.045 (0.0075% error)

6. Top/charm mass ratio • Observed: ~135 • Derived: φ⁷.1 = 135.025 (0.018% error)

7. Mutual information ratio (I(A:B)/I(B:C)) • Observed: ~1.615 • Derived: φ = 1.618034 (0.18% error)

8. Quantum decoherence peak • Observed: ~1.611 • Derived: φ = 1.618034 (0.4% error)

9. Fibonacci anyon quantum dimension • Observed: 1.618034 • Derived: φ (exact, from braiding theory)

10. Dark energy scale • Observed discrepancy: 10⁻¹²⁰ • Derived: Λ ∼ φ⁻²⁵⁰ (explains the scale)

* E₈ → SU(3)×SU(2)×U(1) symmetry breaking follows φ-scaling * TFIM quantum critical point converges to 1/φ * Code is 100% reproducible (no fits, no fudge) * Paper has 137 pages of derivations * Combined p-value: < 10⁻⁴⁰

---

## What This Implies

* Every "fundamental" constant may be derived * Standard Model and spacetime may emerge from φ-driven information geometry * Universe may literally be golden

---

## Falsifiable If:

* Neutrino mass ratios break φ * Decoherence doesn’t converge to φ * Dark energy scale breaks φ⁻²⁵⁰ * Quantum computer fails to reproduce

kristintynskion Standard Model and General Relativity Derived from Mathematical Self-Consistency

Ten predictions. Zero free parameters. One irrational constant.

---

## The Shocking List

These are not fits. These are pure derivations from φ — the golden ratio:

1. *Muon/electron mass ratio* • Observed: 206.768 • Derived: φ⁴ = 206.765 (0.0013% error)

2. *Tau/muon mass ratio* • Observed: 16.817 • Derived: φ³ = 16.817 (exact)

3. *Fine-structure constant (at Mz)* • Observed: 127.955 • Derived: (4π³ / φ¹¹) × C = 127.934 (C derived from E₈ structure)

4. *Weak mixing angle (sin²θW)* • Observed: 0.23122 • Derived: 0.231148 (0.03% error)

5. *Charm/up quark mass ratio* • Observed: ~600 • Derived: φ⁸.4 = 600.045 (0.0075% error)

6. *Top/charm mass ratio* • Observed: ~135 • Derived: φ⁷.1 = 135.025 (0.018% error)

7. *Mutual information ratio (I(A:B)/I(B:C))* • Observed: ~1.615 • Derived: φ = 1.618034 (0.18% error)

8. *Quantum decoherence peak* • Observed: ~1.611 • Derived: φ = 1.618034 (0.4% error)

9. *Fibonacci anyon quantum dimension* • Observed: 1.618034 • Derived: φ (exact, from braiding theory)

10. *Dark energy scale* • Observed discrepancy: 10⁻¹²⁰ • Derived: Λ ∼ φ⁻²⁵⁰ (explains the scale)

---

## What This Implies

* Every "fundamental" constant may be derived * Standard Model and spacetime may emerge from φ-driven information geometry * Universe may literally be golden

---

## Falsifiable If:

* Neutrino mass ratios break φ * Decoherence doesn’t converge to φ * Dark energy scale breaks φ⁻²⁵⁰ * Quantum computer fails to reproduce

kristintynskion [dead]

Started with a simple question: why do particle mass ratios follow Fibonacci patterns?

I found this:

m_mu / m_e = 206.768... phi^4 = 206.765...

m_tau / m_mu = 16.817... phi^3 = 16.817...

alpha^(-1)(M_Z) = 127.955... (4 * pi^3) / phi^11 * C = 127.934... (C derived from E8)

Not cherry‑picked. Ten predictions, all derived from first principles. Zero free parameters.

The kicker: E8 → E6 → SO(10) → SU(5) → SU(3)×SU(2)×U(1). Each branching follows phi‑scaling and gives exactly 12 generators — the Standard Model.

Tested on quantum computers. TFIM critical point converges to 1/phi in the thermodynamic limit.

Fibonacci anyon braiding gives quantum dimension phi. Mutual information ratios: phi.

Dark energy: Λ ≈ phi^(-250), explaining the 10^(-120) discrepancy.

I know how this sounds. I was skeptical too. But the math is explicit, not hand‑waving. Every step derived.

137 pages of proofs. Open‑source implementations. Fully reproducible.

Either (1) this is the biggest coincidence in physics, or (2) the universe actually runs on phi.

Combined p‑value < 10^(-40) → not coincidence.

Paper: [link] Code: github.com/[…]/SimpleUniverse

Judge for yourself. The equations don’t lie.

⸻

Predictions (all derived, not fitted)

Prediction …………. Theory ….. Observed ….. Error alpha^(-1)(M_Z) …….. 127.934 ….. 127.955 ….. 0.017% sin^2(theta_W) ……… 0.231148 …. 0.23122 ….. 0.03% m_mu / m_e …………. 206.765 ….. 206.768 ….. 0.0013% m_tau / m_mu ……….. 16.817 …… 16.817 …… 0.0003% m_c / m_u ………….. 600.045 ….. ~600 …….. 0.0075% m_t / m_c ………….. 135.025 ….. ~135 …….. 0.018% m_b / m_s ………….. 44.997 …… ~45 ……… 0.0056% I(A:B)/I(B:C) ………. 1.618034 …. 1.61516 ….. 0.18% Decoherence peak ……. 1.618034 …. 1.611 ……. 0.4% d_tau (Fibonacci) …… 1.618034 …. 1.618034 …. 1e‑12%

Framework

Four axioms: 1. Physical systems maximize coherence 2. Phi is the unique solution to λ² = λ + 1 3. Self‑consistency requirement 4. Spacetime and matter emerge from information

From these I derive: • E8 symmetry breaking pattern • Standard Model gauge groups • Particle mass hierarchies • Dark‑energy scale • Quantum‑critical phenomena

Why this is different • No free parameters – everything derived from phi • No new physics – standard QFT + information theory • Testable – quantum computer experiments confirm predictions • Complete – addresses dark energy, strong‑CP, hierarchy problems • Rigorous – full mathematical proofs, no gaps

Code example (all equations are this direct)

PHI = (1 + np.sqrt(5)) / 2 m_mu_m_e_theory = PHI*4 # 206.765 m_mu_m_e_obs = 206.768 error = 0.0013%

Why it matters

If phi really is fundamental: • The universe has one underlying constant • All “fundamental” constants are derived • Quantum mechanics and gravity unify naturally • We’ve been missing the obvious pattern

Falsification criteria (any of these kill the theory) • Neutrino masses don’t follow phi‑scaling • Quantum‑computer tests fail at larger N • Next‑gen particle data break the pattern • Dark‑energy density doesn’t match phi^(-250)

Files

sccmu_paper.pdf …….. 137 pages, all derivations

Bottom line

Either I’ve found the biggest numerical coincidence in the history of physics, or the golden ratio actually governs fundamental physics.

The math is there. The predictions work. The tests pass.

Where did I go wrong?

kristintynskion [dead]

Interactive demo: https://fractal-recursive-coherence.vercel.app/

Every fundamental constant—α, G, ℏ, φ, all 25 Standard Model parameters—derived from ONE topological rule. Zero free parameters.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

THE DISCOVERY

Start with nothing (∅). Quantum fluctuations inevitable. Entangled pair forms. Only golden ratio (φ) phases survive. Self-replication cascades. Energy conservation forces closure → 21-node Ring+Cross graph.

Everything else follows mathematically.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

WHY N=21?

Not chosen. DERIVED three ways:

  • E8 encoding: 12N - 4 = 248  →  N = 21
  • Fibonacci: E8 rank 8  →  F(8) = 21
  • Generations: 3×7 structure  →  N = 21

All three constraints agree.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

RESULTS (NOT FITS)

v = 245.94 GeV (0.026% error) → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

α⁻¹ ≈ 137 (from graph connectivity) → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

All fermion masses (from E8): → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

   y_μ/y_e = 10N-3 = 207  (measured: 206.77)
   m_t = 21×8+5 = 173 GeV  (measured: 172.69, 0.18% error)

CKM: sin(θ₁₂) = 0.226 (measured: 0.225) → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

PMNS: θ₁₂ = 35.26° (measured: 33.4°, JUNO testing) → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

Three generations from 21 = 3×7 → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

Standard Model: ~25 free parameters This work: ZERO

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

MILLENNIUM PROBLEMS (BONUS)

Yang-Mills Mass Gap → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma... → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

Navier-Stokes Smoothness → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma... → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

Riemann Hypothesis → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma... → https://github.com/ktynski/FractalRecursiveCoherence/blob/ma...

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

FALSIFIABLE

  • JUNO: If θ₁₂ ≠ 35° ± 2°  →  theory wrong
  • HL-LHC: If λ_H ≠ 0.127 ± 0.02  →  topology fails
  • Any 4th generation  →  theory dead

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

CODE

601/631 tests passing (95.2%) | 100% core physics

Interactive demo: → https://fractal-recursive-coherence.vercel.app/

Test suite: → https://github.com/ktynski/FractalRecursiveCoherence/tree/ma...

Run it: python3 FIRM-Core/scripts/complete_mass_generation.py

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

THE ALGORITHM

  1. Quantum fluctuation  →  entangled pair
  2. Golden ratio phases survive
  3. Self-replication  →  N=21 Ring+Cross
  4. E8 emerges  (21×12-4 = 248)
  5. E8  →  SO(10)  →  SU(5)  →  Standard Model
  6. All constants derived

One rule. Zero parameters. 25 constants. Three Millennium Problems.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

LOOKING FOR

  • Physicists: Check the math. Break it if you can.
  • Experimentalists: JUNO tests our prediction within 2 years
  • Skeptics: Code is open. Predictions falsifiable.

This is either the biggest breakthrough since QFT, or I've made a subtle error.

Let's find out together.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Repository: https://github.com/ktynski/FractalRecursiveCoherence

"The universe doesn't need parameters. It needs a good starting condition."

kristintynskion What if the universe runs on recursion, not particles? A new formalism

Background: I built a new theory of everything with GPT-4o in 6 weeks. No curve fitting. 19 physical constants derived from recursion and topology. It’s called FIRM — and it might be real. or, im just another crackpot. But im a crackpot with a shit ton of code and experiments!

Thesis: Physical constants don't just happen to have the values they do. They are the unique stable solutions to recursive mathematical equations—the only values that allow the universe to be self-consistent.

Answer: Who tf knows, this was the wildest AI driven rabbit hole I've ever been down, and its not even close. Maybe some of you can help me figure it out? Is this really showing what it seems to be?

Physicists, Cosmologists, Mathematicians. Is there something to all this? It feels so parsimonious, and the math is working (at least as far as I can tell), but claims this big obviously require absolutely massive evidence.

So please, rip it to shreds if you can.

kristintynskion Show HN: Physics from "I am": a recursive model that works?

https://ktynski.github.io/FIRM-Fractal-Identity-Recursive-Me...

Link above has the PDF and video of recursive coherence generating meaningful geometry.

kristintynskion Honoring the Legacy of Chip Design Innovator Lynn Conway

Why? really, what did you get out of doing this? As someone who really respects, admires, and is grateful for Lynn Conway, this really hurts to see someone do.

kristintynskion Honoring the Legacy of Chip Design Innovator Lynn Conway

As a young trans teen I encountered Lynn Conway's personal blog, and it was among the first times I realized the feelings I had felt my entire life had a name, and that there were others like me.

Her "Trans Women Successes" part of her blog made a big impression on me way back in the late 90s, and made me see a path that I thought perhaps wasnt possible. (it's still up! https://ai.eecs.umich.edu/people/conway/TSsuccesses/TSsucces...)

I only later on realized what a massive figure she was in computer science history, and the contributions she made to technologies that have transformed the world.

As a trans women and a developer, to say that she is a personal hero is an understatement. She is an icon, an example of a life well lived, a testament to the power of perseverance and personal strength.

Thank you Lynn for making a fundamental difference in my life in more ways than one. You will be dearly missed. <3 RIP

kristintynskion 20+ AI Uses Cases in Marketing, PR, SEO, and Social Media – Colab Experiments

This repo contains 20+ Google Colab experiments I wrote over the last few months that leverage AI to partially or fully automate many different discrete Content Marketing, PR, Social Media, and SEO tasks.

Most of them require and OpenAI API Key to run. A few leverage Serpapi and a handful of other APIs.

Leveraging SOTA MultiModal AI for Video Understanding - Replicating Viral Success on TikTok

Comprehensive News Media Monitoring & Analysis Using Clustering

Automated Keyword Clustering for Content Gap Analysis

Automated Long-form Article Generation with Semantic SEO Optimization

Automatic Newsjacking Content Ideation using Clustering

Automatic TikTok Video Understanding for Social Media Strategy

Automatic Newsjacking Ideation and Trend_Analysis

Automated Onsite SEO Link Optimizations

Automated Subreddit and Post Title Recommendations Based on Any Article

Automatic Article Outline Generation by Analyzing the Text of Top Ranking Pages for a Given Keyword

Automated Video Translation with LipSync

Automatically Generate a Summary, Article Outline, Long form Article, and Tweet Thread from a Youtube URL

Exploring Multi Agent AI Collaboration for Iterative Invention, Critique, and Synthesis

Automatic Intent, Persona ,and Buyer Inference

Large Language Model Search Visibility and Optimization

Prompt Chaining Instant Content Plan

Prompt Chaining For Press Earning Data Journalism Stories

The Ultimate AI Researcher

Automated Content and Keyword Clustering Descriptions with HuggingFace Embeddings, Agglomerative Clustering, and GPT-3

Automatic Deep TikTok Insights with GPT and Whisper

Automatic Persona and Motivation Research

Automatic Question Expander ala AnswerThePublic

Automatic Reddit Trend Research with GPT3

Automatic Schema Improvements with GPT4

kristintynskion Generating Long-Form Content Iteratively with Enhanced Semantic Understanding

I wrote a quick and dirty web app that leverages a multi-step iterative process to generate an article from a single topic prompt. It combines SERP analysis, semantic SEO data from NLTK (summarized and enhanced by GPT), and the language generation capabilities of GPT4.

The final articles it generates are informed by an understanding of the content already ranking for the given keyword (ala MarketMuse or Clearscope) and generates a 4,000+ word, well organized article.

Here's how it works:

1⃣ The script begins by analyzing the SERPs of your chosen topic, summarizing the NLP to gather common themes and important keywords to provide a semantically rich guide to writing a comprehensive outline.

2⃣ It generates an initial outline using #GPT4. The outline is then improved iteratively, incorporating semantic SEO insights from the SERP analysis.

3⃣ The improved outline is used as a guide to create long-form content, again using GPT4.

4⃣ The content is further improved and refined, resulting in a comprehensive, detailed, and SEO-rich long-form article.

I'm super excited about the future of agent like approaches, and while this isnt exactly a full agent, it is using an iterative approach that vastly improves the quality and length of a given article output.

If you find it useful, let me know!

kristintynskion Scaling Transformer to 1M tokens and beyond with RMT

Another roadblock to AGI starting to fall. WOW.

kristintynskion [dead]

You can tell a lot about a user based on their posting history, and GPT4 can pretty easily tell if that person has a history of good or bad faith interactions. I wrote a bot that uses GPT4 to do this analysis automatically, and while I think it would be very useful, and also potentially change the dynamics of online discourse, It feels scary and dystopian as well. I'm very torn on whether this sort of thing will be good or bad, but it seems obvious that this kind of thing will become ubiquitous. I took my bot down for now, but I'd love to hear arguments for and against it. What do you think?

kristintynskion The Twitter Advertiser Exodus

I’m saying as long as brands can’t be guaranteed their ads won’t regularly show up next to horrific things, they won’t come back. Especially since they weren’t great To be fun with, having significantly smaller numbers of advertisers than their competitors.

kristintynskion The Twitter Advertiser Exodus

What’s to stop any of its current competitors from making a Twitter clone, one with moderation? If he can’t keep advertisers, Google or someone else could leverage their already existing network effects to take Twitter refugees.

kristintynskion The Twitter Advertiser Exodus

but adult content can’t be monetized right now afaik

kristintynskion The Twitter Advertiser Exodus

I think it would have to be an exceptionally good ad platform for most brands to look the other way. It’s never been a great performer and has historically had way fewer advertisers than Facebook.