AI Artificial Intelligence
Artificial Intelligence
Future thinking.
We staan aan de wieg van AI op dit moment. Het leuk wat we er mee kunnen en er zijn nuttige toepassingen die nu al gebruikt worden met AI. Denk aan diagnostisch werk in ziekenhuizen.
Maar echt het volle potentieel van AI is nog niet bereikt naar mijn mening.
En dat zal sneller gaan dan we mogelijk willen...
Want wat gebeurd er als het mogelijk is om AI te "draaien" op quantum computers.
Welke vooruitgang staat ons dan te wachten.
We zitten nu in het NISQ (Noisy Intermediate-Scale Quantum) stuk van de ontwikkeling. Nog teveel foutgevoelig en te klein om hier serieus iets mee te gaan doen op grote schaal.
Dat gaat veranderen als we Fault-Tolerant Quantum Computing (FTQC) bereiken. Volgens experts kan het nog een decennium duren voordat we dat bereiken.
Maar...... AI is nu al aan het helpen met quantum computing. Dus de kans is aanwezig dat er een versnelling zal plaats vinden om eerder FTQC te bereiken.
Hieronder een stuk vanuit Gemini:
The relationship between AI and quantum computing isn't a one-way street; it is a powerful feedback loop. AI is actively acting as a catalyst, compressing what used to be a rigid 10-to-15-year roadmap into a significantly shorter timeline.
In fact, recent breakthroughs have caused major tech companies and cybersecurity experts to radically pull forward their timelines.
Here is exactly how AI is shrinking the runway to Fault-Tolerant Quantum Computing (FTQC):
1. Solving the Error Correction Overhead
The biggest barrier to FTQC has always been the "overhead problem." To get just 1 stable, error-corrected logical qubit, you historically needed thousands of messy, unstable physical qubits just to watch for errors.
• The AI Acceleration: Scientists are using AI as an active discovery engine to explore algorithmic search spaces. In late 2025, teams at places like QuEra introduced AI-driven Algorithmic Fault Tolerance (AFT), which reduced the computational effort spent on error correction by up to 100 times.
• The Result: Instead of needing millions of physical qubits, AI-assisted hardware optimizations suggest that 10,000 to 26,000 qubits might soon be enough to achieve highly advanced, crypto-graphically relevant computing.
2. Extreme Material and Hardware Calibration
Quantum chips are incredibly sensitive. Slight changes in temperature, magnetic fields, or manufacturing flaws can ruin a calculation.
• Humans cannot manually calibrate thousands of interconnected qubits in real time.
• Advanced machine learning models are now used to manage real-time error mitigation, hardware calibration, and material discovery. AI can predict and counteract "noise" on a quantum chip before it even happens, effectively forcing imperfect hardware to act like stable, fault-tolerant hardware.
3. Crashing Timelines (The Proof)
Because AI has become quantum's "co-pilot," the timelines aren't just theoretical anymore—they are shifting rapidly:
• The 2029 Shift: Major players like IBM and Google Quantum AI have officially aligned their roadmaps to deploy practical, fault-tolerant quantum systems by 2029.
• Urgent Preparations: In early 2026, breakthroughs in AI-assisted quantum algorithm compilation sent a jolt through cybersecurity. Infrastructure companies like Cloudflare immediately bumped their "quantum-readiness" deadlines forward to 2029 because quantum computers capable of cracking encryption are suddenly arriving years earlier than expected.
The Ultimate Loop
We are entering a compounding cycle: AI accelerates the building of quantum computers -> Quantum computers run next-generation AI -> Next-generation AI builds even better quantum computers.
Because of this compounding effect, we are moving out of the "theoretical math" phase and directly into the engineering phase.
En over het AI op quantum computing stuk:
the integration of AI and quantum computing has the potential to drive a monumental leap forward, completely redefining what artificial intelligence can achieve. While today's AI relies on brute-forcing massive amounts of data through classical silicon chips, quantum-powered AI changes the core mathematics of how computers think.
Here is how quantum computing will push AI far beyond its current limits:
1. Cracking "Exponentially Hard" Problems
Classical AI struggles with combinatorial explosion—problems where the number of possible choices grows exponentially (like simulating molecular bonds, logistics optimization, or predicting global climate patterns).
• Current AI: Tries to approximate answers because calculating every option takes too long.
• Quantum AI: Evaluates massive numbers of possibilities simultaneously due to superposition, finding optimal solutions in minutes that would take classical supercomputers millennia.
2. A Leap Beyond Next-Token Prediction
Modern AI (like large language models) doesn't truly understand concepts; it excels at predicting the most likely next word or pixel based on statistical patterns.
• Quantum mechanics natively handles complex, interconnected correlation patterns (entanglement).
• This allows quantum AI to recognize intricate, hidden relationships in data that classical networks physically cannot see, potentially moving us from simple pattern-matching toward genuine reasoning and abstract understanding.
3. Infinite Scalability Without Power Grids
Right now, scaling AI requires building massive data centers that demand as much electricity as small cities. Quantum systems utilize quantum states rather than pushing billions of electrons through transistors. Once stabilized, quantum AI could train vastly more sophisticated models using a fraction of the energy required by today's GPU clusters.
4. Overcoming the "Data Wall"
Current AI models are running out of high-quality human data to train on. Quantum AI doesn't just process existing data faster; it is uniquely suited to accurately simulate complex environments (like quantum physics or biochemistry). This allows the AI to generate its own highly accurate, synthetic training data to learn things humanity hasn't even discovered yet.
Gaat dit voor ons een "blessing" worden of een serieuze nachtmerrie?
We staan aan de wieg van AI op dit moment. Het leuk wat we er mee kunnen en er zijn nuttige toepassingen die nu al gebruikt worden met AI. Denk aan diagnostisch werk in ziekenhuizen.
Maar echt het volle potentieel van AI is nog niet bereikt naar mijn mening.
En dat zal sneller gaan dan we mogelijk willen...
Want wat gebeurd er als het mogelijk is om AI te "draaien" op quantum computers.
Welke vooruitgang staat ons dan te wachten.
We zitten nu in het NISQ (Noisy Intermediate-Scale Quantum) stuk van de ontwikkeling. Nog teveel foutgevoelig en te klein om hier serieus iets mee te gaan doen op grote schaal.
Dat gaat veranderen als we Fault-Tolerant Quantum Computing (FTQC) bereiken. Volgens experts kan het nog een decennium duren voordat we dat bereiken.
Maar...... AI is nu al aan het helpen met quantum computing. Dus de kans is aanwezig dat er een versnelling zal plaats vinden om eerder FTQC te bereiken.
Hieronder een stuk vanuit Gemini:
The relationship between AI and quantum computing isn't a one-way street; it is a powerful feedback loop. AI is actively acting as a catalyst, compressing what used to be a rigid 10-to-15-year roadmap into a significantly shorter timeline.
In fact, recent breakthroughs have caused major tech companies and cybersecurity experts to radically pull forward their timelines.
Here is exactly how AI is shrinking the runway to Fault-Tolerant Quantum Computing (FTQC):
1. Solving the Error Correction Overhead
The biggest barrier to FTQC has always been the "overhead problem." To get just 1 stable, error-corrected logical qubit, you historically needed thousands of messy, unstable physical qubits just to watch for errors.
• The AI Acceleration: Scientists are using AI as an active discovery engine to explore algorithmic search spaces. In late 2025, teams at places like QuEra introduced AI-driven Algorithmic Fault Tolerance (AFT), which reduced the computational effort spent on error correction by up to 100 times.
• The Result: Instead of needing millions of physical qubits, AI-assisted hardware optimizations suggest that 10,000 to 26,000 qubits might soon be enough to achieve highly advanced, crypto-graphically relevant computing.
2. Extreme Material and Hardware Calibration
Quantum chips are incredibly sensitive. Slight changes in temperature, magnetic fields, or manufacturing flaws can ruin a calculation.
• Humans cannot manually calibrate thousands of interconnected qubits in real time.
• Advanced machine learning models are now used to manage real-time error mitigation, hardware calibration, and material discovery. AI can predict and counteract "noise" on a quantum chip before it even happens, effectively forcing imperfect hardware to act like stable, fault-tolerant hardware.
3. Crashing Timelines (The Proof)
Because AI has become quantum's "co-pilot," the timelines aren't just theoretical anymore—they are shifting rapidly:
• The 2029 Shift: Major players like IBM and Google Quantum AI have officially aligned their roadmaps to deploy practical, fault-tolerant quantum systems by 2029.
• Urgent Preparations: In early 2026, breakthroughs in AI-assisted quantum algorithm compilation sent a jolt through cybersecurity. Infrastructure companies like Cloudflare immediately bumped their "quantum-readiness" deadlines forward to 2029 because quantum computers capable of cracking encryption are suddenly arriving years earlier than expected.
The Ultimate Loop
We are entering a compounding cycle: AI accelerates the building of quantum computers -> Quantum computers run next-generation AI -> Next-generation AI builds even better quantum computers.
Because of this compounding effect, we are moving out of the "theoretical math" phase and directly into the engineering phase.
En over het AI op quantum computing stuk:
the integration of AI and quantum computing has the potential to drive a monumental leap forward, completely redefining what artificial intelligence can achieve. While today's AI relies on brute-forcing massive amounts of data through classical silicon chips, quantum-powered AI changes the core mathematics of how computers think.
Here is how quantum computing will push AI far beyond its current limits:
1. Cracking "Exponentially Hard" Problems
Classical AI struggles with combinatorial explosion—problems where the number of possible choices grows exponentially (like simulating molecular bonds, logistics optimization, or predicting global climate patterns).
• Current AI: Tries to approximate answers because calculating every option takes too long.
• Quantum AI: Evaluates massive numbers of possibilities simultaneously due to superposition, finding optimal solutions in minutes that would take classical supercomputers millennia.
2. A Leap Beyond Next-Token Prediction
Modern AI (like large language models) doesn't truly understand concepts; it excels at predicting the most likely next word or pixel based on statistical patterns.
• Quantum mechanics natively handles complex, interconnected correlation patterns (entanglement).
• This allows quantum AI to recognize intricate, hidden relationships in data that classical networks physically cannot see, potentially moving us from simple pattern-matching toward genuine reasoning and abstract understanding.
3. Infinite Scalability Without Power Grids
Right now, scaling AI requires building massive data centers that demand as much electricity as small cities. Quantum systems utilize quantum states rather than pushing billions of electrons through transistors. Once stabilized, quantum AI could train vastly more sophisticated models using a fraction of the energy required by today's GPU clusters.
4. Overcoming the "Data Wall"
Current AI models are running out of high-quality human data to train on. Quantum AI doesn't just process existing data faster; it is uniquely suited to accurately simulate complex environments (like quantum physics or biochemistry). This allows the AI to generate its own highly accurate, synthetic training data to learn things humanity hasn't even discovered yet.
Gaat dit voor ons een "blessing" worden of een serieuze nachtmerrie?
OxygeneFRL-vrijdag 8 mei 2020 @ 08:52:59: Ik had een pleuris hekel aan je maar nu ik weet dat je tegen een vuurwerkverbod ben, hou ik van je.
Zodra computers slimmer zijn dan mensen kunnen die computers ook sneller ontwikkelen, ik meen dat we dat punt vorig jaar bereikt hadden
Op dinsdag 25 februari 2020 12:55 schreef Pumpalov het volgende:
een beetje zo'n lichtgetinte inteeltkop als dat filosoofert heeft.
een beetje zo'n lichtgetinte inteeltkop als dat filosoofert heeft.
|
|
