JasonC
@jasonfromatlanta.bsky.social
I am someone that likes puzzles...I am always working one. I am someone that connects things across domains. I see patterns.
Indeed. It exposes every single control surface along the cascade. Shows you exactly where and how to intervene. The only question is wheither or not you can reach the control....but it can tell you that too.
February 6, 2026 at 9:46 PM
Indeed. It exposes every single control surface along the cascade. Shows you exactly where and how to intervene. The only question is wheither or not you can reach the control....but it can tell you that too.
The fold is everything.
February 6, 2026 at 9:26 PM
The fold is everything.
What a $50 Trillion Market and a Soap Bubble Have in Common? The Option Price acts like Soap Film.
Black-Scholes earned a Nobel Prize, and has determined stock option price since. It shares roots with the heat equation, but did you know both have a geometric root?
medium.com/p/what-a-50-...
Black-Scholes earned a Nobel Prize, and has determined stock option price since. It shares roots with the heat equation, but did you know both have a geometric root?
medium.com/p/what-a-50-...
What a $50 Trillion Market and a Soap Bubble Have in Common
The Option Price Is a Soap Film
medium.com
February 5, 2026 at 7:47 PM
What a $50 Trillion Market and a Soap Bubble Have in Common? The Option Price acts like Soap Film.
Black-Scholes earned a Nobel Prize, and has determined stock option price since. It shares roots with the heat equation, but did you know both have a geometric root?
medium.com/p/what-a-50-...
Black-Scholes earned a Nobel Prize, and has determined stock option price since. It shares roots with the heat equation, but did you know both have a geometric root?
medium.com/p/what-a-50-...
Reposted by JasonC
Paper demonstrates hat musical consonance is a a quantifiable measure of computational efficiency on a finite-capacity scaffold.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
February 4, 2026 at 1:56 PM
Paper demonstrates hat musical consonance is a a quantifiable measure of computational efficiency on a finite-capacity scaffold.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
Reposted by JasonC
Read “The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Time Bomb“ by Jason Connerty on Medium: medium.com/@jasonrconne...
The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Time Bomb
Jason Connerty | January 2026
medium.com
January 25, 2026 at 2:11 AM
Read “The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Time Bomb“ by Jason Connerty on Medium: medium.com/@jasonrconne...
The Hidden Geometry of Claude’s Constitution.
Anthropic’s Constitution quietly reflects ideas that show up across physics and complex systems.. phase transitions, boundary dynamics, etc.
Here we explore what happens when you treat AI alignment as a geometry problem..
medium.com/p/the-hidden...
Anthropic’s Constitution quietly reflects ideas that show up across physics and complex systems.. phase transitions, boundary dynamics, etc.
Here we explore what happens when you treat AI alignment as a geometry problem..
medium.com/p/the-hidden...
The Hidden Geometry of Claude’s Constitution
A Constraint-Based Analysis of Anthropic’s Alignment Framework
medium.com
February 5, 2026 at 3:55 PM
The Hidden Geometry of Claude’s Constitution.
Anthropic’s Constitution quietly reflects ideas that show up across physics and complex systems.. phase transitions, boundary dynamics, etc.
Here we explore what happens when you treat AI alignment as a geometry problem..
medium.com/p/the-hidden...
Anthropic’s Constitution quietly reflects ideas that show up across physics and complex systems.. phase transitions, boundary dynamics, etc.
Here we explore what happens when you treat AI alignment as a geometry problem..
medium.com/p/the-hidden...
Paper demonstrates hat musical consonance is a a quantifiable measure of computational efficiency on a finite-capacity scaffold.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
February 4, 2026 at 1:56 PM
Paper demonstrates hat musical consonance is a a quantifiable measure of computational efficiency on a finite-capacity scaffold.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
By modeling the auditory system as a ledger processing distinct identities (frequencies), we derive the hierarchy of harmonic intervals.
A 2,500 year old music theory question just found its root.
Why do some note combinations sound stable and others sound tense?
The answer is geometric, and it's been hiding inside the circle of fifths the whole time.
I wrote the accessible version here: medium.com/@jasonrconne...
Why do some note combinations sound stable and others sound tense?
The answer is geometric, and it's been hiding inside the circle of fifths the whole time.
I wrote the accessible version here: medium.com/@jasonrconne...
For the Love of Music: The Geometry of Dissonance
A 2,500 Year Old Music Theory Question Gets a Shape
medium.com
February 4, 2026 at 12:59 PM
A 2,500 year old music theory question just found its root.
Why do some note combinations sound stable and others sound tense?
The answer is geometric, and it's been hiding inside the circle of fifths the whole time.
I wrote the accessible version here: medium.com/@jasonrconne...
Why do some note combinations sound stable and others sound tense?
The answer is geometric, and it's been hiding inside the circle of fifths the whole time.
I wrote the accessible version here: medium.com/@jasonrconne...
Why complex numbers?
because you need them for unistochastic matrices when N > 2.
But there's a deeper question: why does nature instantiate indivisible processes at all?
If reality has finite distinguishability... a fundamental resolution limit... then indivisibility isn't arbitrary.
because you need them for unistochastic matrices when N > 2.
But there's a deeper question: why does nature instantiate indivisible processes at all?
If reality has finite distinguishability... a fundamental resolution limit... then indivisibility isn't arbitrary.
New on the Archive:
Barandes, Jacob A. (2026) A Deflationary Account of Quantum Theory and its Implications for the Complex Numbers. [Preprint]
https://philsci-archive.pitt.edu/28062/
Barandes, Jacob A. (2026) A Deflationary Account of Quantum Theory and its Implications for the Complex Numbers. [Preprint]
https://philsci-archive.pitt.edu/28062/
February 1, 2026 at 5:19 PM
Why complex numbers?
because you need them for unistochastic matrices when N > 2.
But there's a deeper question: why does nature instantiate indivisible processes at all?
If reality has finite distinguishability... a fundamental resolution limit... then indivisibility isn't arbitrary.
because you need them for unistochastic matrices when N > 2.
But there's a deeper question: why does nature instantiate indivisible processes at all?
If reality has finite distinguishability... a fundamental resolution limit... then indivisibility isn't arbitrary.
Derived Navier-Stokes equations as theorem from 2 constraints; no empirics.
Turbulence as resolution saturation (advection > diffusion → fractal eddies). Ties to cosmic web.
PDF: zenodo.org/records/1844...
#Math #Physics #Turbulence #MillenniumPrize
Turbulence as resolution saturation (advection > diffusion → fractal eddies). Ties to cosmic web.
PDF: zenodo.org/records/1844...
#Math #Physics #Turbulence #MillenniumPrize
February 1, 2026 at 4:16 PM
Derived Navier-Stokes equations as theorem from 2 constraints; no empirics.
Turbulence as resolution saturation (advection > diffusion → fractal eddies). Ties to cosmic web.
PDF: zenodo.org/records/1844...
#Math #Physics #Turbulence #MillenniumPrize
Turbulence as resolution saturation (advection > diffusion → fractal eddies). Ties to cosmic web.
PDF: zenodo.org/records/1844...
#Math #Physics #Turbulence #MillenniumPrize
I just derived the dark matter ratio from topology. No free parameters. 5.32:1 predicted, 5.36:1 observed. <1% error.
Two axioms → 21 layers → matter, force, gravity, entropy, relativity, quantum mechanics, dark matter, dark energy, black holes, Big Bang.
Two axioms → 21 layers → matter, force, gravity, entropy, relativity, quantum mechanics, dark matter, dark energy, black holes, Big Bang.
January 31, 2026 at 12:46 AM
I just derived the dark matter ratio from topology. No free parameters. 5.32:1 predicted, 5.36:1 observed. <1% error.
Two axioms → 21 layers → matter, force, gravity, entropy, relativity, quantum mechanics, dark matter, dark energy, black holes, Big Bang.
Two axioms → 21 layers → matter, force, gravity, entropy, relativity, quantum mechanics, dark matter, dark energy, black holes, Big Bang.
Read “The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Time Bomb“ by Jason Connerty on Medium: medium.com/@jasonrconne...
The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Time Bomb
Jason Connerty | January 2026
medium.com
January 25, 2026 at 2:11 AM
Read “The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Time Bomb“ by Jason Connerty on Medium: medium.com/@jasonrconne...
Do you know how to measure for fragility within an organization? It is both predictable and preventable.
The article isn't a rehash of the Verizon outage. No. It looks at the true root cause (hint..it isn't the bad update), then walks through the "how to" of building an early warning system.
The article isn't a rehash of the Verizon outage. No. It looks at the true root cause (hint..it isn't the bad update), then walks through the "how to" of building an early warning system.
Verizon's outage last week made me think..
When you optimize by collapsing your network to one solution while purging the people who knew the workarounds, you flatten to a 1D line. First shock shatters everything.
The signals are always there before the snap.
medium.com/@jasonrconne...
When you optimize by collapsing your network to one solution while purging the people who knew the workarounds, you flatten to a 1D line. First shock shatters everything.
The signals are always there before the snap.
medium.com/@jasonrconne...
The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Clock
Jason Connerty | January 2026
medium.com
January 19, 2026 at 8:00 PM
Do you know how to measure for fragility within an organization? It is both predictable and preventable.
The article isn't a rehash of the Verizon outage. No. It looks at the true root cause (hint..it isn't the bad update), then walks through the "how to" of building an early warning system.
The article isn't a rehash of the Verizon outage. No. It looks at the true root cause (hint..it isn't the bad update), then walks through the "how to" of building an early warning system.
Reposted by JasonC
Prompt injection is a structural property of how we process information. Until we harden the resolution boundary through architectural change, the cycle continues .
Full analysis here: medium.com/@jasonrconne...
Full analysis here: medium.com/@jasonrconne...
Breaking AI’s Vicious Security Cycle: Why AI Security Guardrails Keep Failing
A Constraint-Based Analysis of Prompt Injection
medium.com
January 11, 2026 at 6:30 PM
Prompt injection is a structural property of how we process information. Until we harden the resolution boundary through architectural change, the cycle continues .
Full analysis here: medium.com/@jasonrconne...
Full analysis here: medium.com/@jasonrconne...
Reposted by JasonC
Quick add: This maps closely to 'critical slowing down' used as EWS in ecology/climate (e.g., PNAS papers on DL detecting tipping points). Training is just another dynamical system...why wouldn't it show the same?
AI training collapses aren’t random; they resemble the “critical slowing down” seen before physical phase transitions. That’s an opportunity.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
January 13, 2026 at 5:20 PM
Quick add: This maps closely to 'critical slowing down' used as EWS in ecology/climate (e.g., PNAS papers on DL detecting tipping points). Training is just another dynamical system...why wouldn't it show the same?
Reposted by JasonC
Wrote an AI race analysis in September. Checked back in four months later.
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
The AI Race Scorecard: A Four-Month Check-In
China’s AI Momentum Accelerates as U.S. Procurement Prioritizes Politics Over Power
medium.com
January 15, 2026 at 12:23 PM
Wrote an AI race analysis in September. Checked back in four months later.
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
Verizon's outage last week made me think..
When you optimize by collapsing your network to one solution while purging the people who knew the workarounds, you flatten to a 1D line. First shock shatters everything.
The signals are always there before the snap.
medium.com/@jasonrconne...
When you optimize by collapsing your network to one solution while purging the people who knew the workarounds, you flatten to a 1D line. First shock shatters everything.
The signals are always there before the snap.
medium.com/@jasonrconne...
The Geometry of Fragility: Why Your Optimization Strategy Is a Ticking Clock
Jason Connerty | January 2026
medium.com
January 19, 2026 at 5:45 PM
Verizon's outage last week made me think..
When you optimize by collapsing your network to one solution while purging the people who knew the workarounds, you flatten to a 1D line. First shock shatters everything.
The signals are always there before the snap.
medium.com/@jasonrconne...
When you optimize by collapsing your network to one solution while purging the people who knew the workarounds, you flatten to a 1D line. First shock shatters everything.
The signals are always there before the snap.
medium.com/@jasonrconne...
Wrote an AI race analysis in September. Checked back in four months later.
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
The AI Race Scorecard: A Four-Month Check-In
China’s AI Momentum Accelerates as U.S. Procurement Prioritizes Politics Over Power
medium.com
January 15, 2026 at 12:23 PM
Wrote an AI race analysis in September. Checked back in four months later.
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
The Short: China validated hardware sovereignty at scale. The US announced classified network access for an AI under investigation for kid porn, because it "won't be woke."
I was too optimistic.
medium.com/@jasonrconne...
Back in September I wrote a piece on the US-China AI race, flagging this risk: "When procurement embeds political constraints into capability requirements, optimization migrates into the design layer."
January 14, 2026 at 1:43 PM
Back in September I wrote a piece on the US-China AI race, flagging this risk: "When procurement embeds political constraints into capability requirements, optimization migrates into the design layer."
Quick add: This maps closely to 'critical slowing down' used as EWS in ecology/climate (e.g., PNAS papers on DL detecting tipping points). Training is just another dynamical system...why wouldn't it show the same?
AI training collapses aren’t random; they resemble the “critical slowing down” seen before physical phase transitions. That’s an opportunity.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
January 13, 2026 at 5:20 PM
Quick add: This maps closely to 'critical slowing down' used as EWS in ecology/climate (e.g., PNAS papers on DL detecting tipping points). Training is just another dynamical system...why wouldn't it show the same?
AI training collapses aren’t random; they resemble the “critical slowing down” seen before physical phase transitions. That’s an opportunity.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
January 13, 2026 at 4:26 PM
AI training collapses aren’t random; they resemble the “critical slowing down” seen before physical phase transitions. That’s an opportunity.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
Spot the signal early: recovery from tiny perturbations slows well before loss explodes. Collapse can be predicted...and sometimes prevented.
The industry keeps building better locks on the exits while attackers are getting in through the lobby.
The system decides what to trust before it decides what to do. Current defenses intervene after that decision. That's why they always fail.
Open to discussions.. Atlanta-based, remote-friendly.
The system decides what to trust before it decides what to do. Current defenses intervene after that decision. That's why they always fail.
Open to discussions.. Atlanta-based, remote-friendly.
January 12, 2026 at 3:33 PM
The industry keeps building better locks on the exits while attackers are getting in through the lobby.
The system decides what to trust before it decides what to do. Current defenses intervene after that decision. That's why they always fail.
Open to discussions.. Atlanta-based, remote-friendly.
The system decides what to trust before it decides what to do. Current defenses intervene after that decision. That's why they always fail.
Open to discussions.. Atlanta-based, remote-friendly.
We’re stuck in a vicious cycle of AI security: vulnerability found -> guardrail deployed -> bypass discovered -> repeat .
The reason? We are trying to solve a hardware problem with software patches. Current "guardrails" are Tier 3 interventions...the lowest leverage possible.
The reason? We are trying to solve a hardware problem with software patches. Current "guardrails" are Tier 3 interventions...the lowest leverage possible.
January 11, 2026 at 6:25 PM
We’re stuck in a vicious cycle of AI security: vulnerability found -> guardrail deployed -> bypass discovered -> repeat .
The reason? We are trying to solve a hardware problem with software patches. Current "guardrails" are Tier 3 interventions...the lowest leverage possible.
The reason? We are trying to solve a hardware problem with software patches. Current "guardrails" are Tier 3 interventions...the lowest leverage possible.
A new article that explores our universe's beginnings, the big bang, and 2D space...I would love to hear your thoughts.
medium.com/@jasonrconne...
medium.com/@jasonrconne...
The Dawn of Reality: Did the Universe Begin with Two Points and a Twist?
A Journey from Observation to Origin
medium.com
December 30, 2025 at 7:29 PM
A new article that explores our universe's beginnings, the big bang, and 2D space...I would love to hear your thoughts.
medium.com/@jasonrconne...
medium.com/@jasonrconne...
So incredibly short sighted..it is disgusting to watch this happen.
The Trump administration have gutted science agencies, terminated research programmes and cancelled billions of dollars in grants to universities
go.nature.com/4pMeAuU
go.nature.com/4pMeAuU
Will US science survive Trump 2.0?
President Donald Trump and his administration have gutted science agencies, terminated research programmes and cancelled billions of dollars in grants to universities. What are the long-term impacts for the United States and the world?
go.nature.com
December 27, 2025 at 4:32 PM
So incredibly short sighted..it is disgusting to watch this happen.