Independent Interpretability Researcher and Runtime Architecture
Joined August 2025
- Tweets 624
- Following 42
- Followers 81
- Likes 43
145 Photos and videos
Pinned Tweet
The missing piece in the MIT model isn’t a mechanism—it’s a cascade. The ‘objective friction’ doesn’t disappear all at once. It’s gradually removed across compression, alignment, UX, and interaction loops
2
108
Who cares fuck elon
Elon Musk is now a trillionaire, and 4,400 SpaceX employees are multimillionaires after the record-breaking IPO
Watch The DeVory Darkins Show on OBBM Network TV — add us to your Roku, FireTV or AppleTV today!
#OBBMNetwork #OBBMNetworkTV
5
open question. literally ANYONE can answer. please. defend elon musk's hoard.
can you tell me what his 'wealth' is grounded in when it comes to valuation? specifically:
Who audits the balance sheets for the companies he is in charge of?
Who sets the prices those shares/companies are worth?
Who monitors the receivables from those audits?
If the answers to any of those is "Tesla"."xAI" or "SpaceX", that simply is not okay.
At what point does valuation become self-referential?
Those are also questions that become increasingly relevant as valuations move away from traditional accounting metrics and toward future expectations, narratives, network effects, AI promises, founder reputation, and market psychology.
@grok, care to take a crack at grounding this? i almost guarantee you cant.
1
1
186
so apparently i've received an adjunct promotion to assistant social media manager as well as chief governance officer for my startup.
how does one manage social media for a budding company?
by doing the same thing i did to learn AI: use chatgpt, claude, grok, and gemini to learn the skill.
not do the skill.
learn the skill.
1
29
Anthropic's Project Glasswing got us thinking about a question that we don't see discussed very often:
What happens when a weak security program gets access to a very powerful security tool?
Most conversations about AI in cybersecurity focus on capability.
Faster analysis. More visibility. More findings. More automation.
Those are all valuable.
But capability and effectiveness aren't the same thing.
Imagine an organization with limited security maturity.
Asset inventories are incomplete. Ownership is unclear. Patch management is inconsistent. Risk prioritization is largely reactive.
Now give that organization an AI system capable of identifying hundreds of potential issues, anomalies, and indicators.
What happens?
The intuitive answer is that security improves.
The practical answer is more complicated.
If an AI system surfaces 400 findings, who decides which 10 matter most?
What criteria are they using? Who owns remediation? How do those findings map to business risk and compliance baselines?
At some point, the bottleneck stops being detection and becomes decision-making.
The organization doesn't have a visibility problem anymore.
It has a governance problem.
The cybersecurity industry has spent decades improving our ability to discover things. The next challenge is deciding what deserves attention, what can wait, and what actions can be verified and trusted under strict compliance standards.
This isn't a theoretical concern for us.
Chelys was built around exactly this gap: helping organizations turn security signals into governed, verifiable action.
Our platform is designed around deterministic guardrails, structured compliance, and verified execution because better detection only matters when teams can trust what happens next.
Our site is officially live.
chelys.io/
#Cybersecurity #Compliance #SecOps #Governance #RiskManagement
1
41
CHATGPT X TACO BELL JUST DROPPED!!!
linkedin.com/pulse/prelimina…
8
🚨 EXCITING OPPORTUNITY ALERT 🚨
After several quarters of strategic alignment, stakeholder engagement, and complete financial collapse, we are thrilled to announce that Vibe to Cognitive Infrastructure Transformation is actively seeking a:
💰 Chief Financial Officer (CFO) 💰
This is a unique opportunity to join a high-growth organization operating at the intersection of:
✨ Cognitive Infrastructure
✨ AI-Native Synergy
✨ Vibe-Driven Capital Allocation
✨ Premium Feline Logistics
Key responsibilities include:
📈 Liquidity Management
🐈 Stabilization of Kitty Purry's premium kibble supply chain
🧾 Review of transaction records consisting primarily of fast-food receipts and hardware invoices
🌈 Cross-functional vibe auditing
🧠 Strategic guidance during late-night architectural discussions
Compensation package includes:
💵 $0.00 salary
📊 Equity of the soul
🚀 Exposure to cutting-edge haunted bullshit
🎬 Performance-based cinematic vibelets
⭐ Immaculate, unregulated vibes
Ideal candidates should possess:
✅ Immunity to corporate embarrassment
✅ Experience operating under uncertain financial realities
✅ Comfort with credit scores influenced by lunar activity
✅ A deep appreciation for emerging cognitive infrastructure paradigms
We are particularly interested in candidates who understand that attention is the most valuable asset class of the 21st century.
If this sounds like the next step in your professional journey, we encourage you to apply immediately.
#Hiring #CFO #Finance #Leadership #AI #Innovation #DigitalTransformation #FutureOfWork #CognitiveInfrastructure #VibeDrivenDevelopment #PatronOfTheArts #BuildingHauntedBullshit
lnkd.in/eRhGkt9C
18
"take me out to the ballgame" is music related to baseball, which fits the hint they gave.
The emojis are infinitely confusing for a model to parse with current technology in relation to tokenization and embedding, and this should NOT be considered a bar for intelligence. It's a limitation of the current technology.
This is why human judgement is important.
AI is not going to take your job
the most effective individuals understand the limitations.
yeah, AI successfully accomplishes A LOT of tasks.
moreso if you build and ground your system correctly.
the successful tasks aren't important though. Those don't move the needle in any direction.
The only perception that has an effect on trust, reliably, and adoption are documented production failures.
the trust curve for use has to outweigh the potential blast radius for error in order to see the technology adopted SUCCESSFULLY. Right now, that isn't even close
12
Check out my latest article: here's something to laugh at. hope y'all enjoy it. linkedin.com/pulse/heres-som… via @LinkedIn
5
May 16
you guys know that openclaw thing you hear about?
i gave it governance for free.
enjoy.
github.com/Bradsadevnow/claw…
13
May 15
here at #ApertureLaboratories, we'd like to reassure everyone that we have heard the concerns in regard to the mistreatment of our test subjects.
we plan to spearhead a new internal initiative to improve mandatory safety compliance.
rest assured, this price will be passed on through the wide family of #Aperture products to the consumer.
Thank you for your participation in this sweeping reform.
9
May 14
I could use your help with my fundraiser on GoFundMe. Please share, support, or donate—every small action counts. gofund.me/69c5e7c72
2
May 14
i'd like to share communication that i got from someone that stresses WHY shared language is important.
it's important to note that this person sent me a test harness.
unfortunately, because anyone can be an engineer, they don't understand what a test harnass is.
combine that with the fact that they think i stole their language because i was using the same language they did.
this may become a legal issue i legit cannot afford, and i genuinely have done nothing besides create my own framework.
not to mention i have to defend my honor, integrity, and legitimate research.
4