Finding the Next Aster: 5 High-Revenue, Un-Tokenized Perp DEXs
This article spotlights five high-revenue, yet un-tokenized Decentralized Perpetual Exchanges (Perp DEXs), focusing on their protocol revenue, technical features, and growth potential. These projects demonstrate genuine profitability amidst intense competition in the sector. edgeX: The High-Performance Contender edgeX set a new revenue record for Perp DEXs in September 2025, with cumulative revenue reaching $49.47 million, solidifying its position as the second-highest revenue generator in th...
Can PoL v2 Ignite a BeraChain Rally?
1. Core Breakthrough: From Mercenary Liquidity to Value Feedback Loop In a post-yield-farming world, the only question that matters is “how does a chain manufacture its own organic demand?” Berachain’s answer is to make the native token the first beneficiary of every unit of growth. Proof-of-Liquidity (PoL) v2 flips the old script. Instead of letting ETH/SOL-style gas tokens watch from the sidelines while DeFi protocols pocket the upside, v2 reroutes 33 % of all DApp-bribe incentives from BGT...
Can DeepSeek Stay Hot?
DeepSeek is set to face more pressure and challenges in the future. The race towards a universal model has just begun, and who will ultimately win depends on continuous investment in funding and technological iteration. · A headhunter responsible for sourcing high-end tech talent in the large model field told The Paper Technology that DeepSeek's hiring logic is not much different from that of other companies in the large model sector. The core label for talent is "young and high-potential," m...
Support Riley
Ten Hard Truths About Today’s AI: It Won’t Steal Your Job—It Just Turns You into a Passable “Workhorse”
Support Riley
Ten Hard Truths About Today’s AI: It Won’t Steal Your Job—It Just Turns You into a Passable “Workhorse”
Finding the Next Aster: 5 High-Revenue, Un-Tokenized Perp DEXs
This article spotlights five high-revenue, yet un-tokenized Decentralized Perpetual Exchanges (Perp DEXs), focusing on their protocol revenue, technical features, and growth potential. These projects demonstrate genuine profitability amidst intense competition in the sector. edgeX: The High-Performance Contender edgeX set a new revenue record for Perp DEXs in September 2025, with cumulative revenue reaching $49.47 million, solidifying its position as the second-highest revenue generator in th...
Can PoL v2 Ignite a BeraChain Rally?
1. Core Breakthrough: From Mercenary Liquidity to Value Feedback Loop In a post-yield-farming world, the only question that matters is “how does a chain manufacture its own organic demand?” Berachain’s answer is to make the native token the first beneficiary of every unit of growth. Proof-of-Liquidity (PoL) v2 flips the old script. Instead of letting ETH/SOL-style gas tokens watch from the sidelines while DeFi protocols pocket the upside, v2 reroutes 33 % of all DApp-bribe incentives from BGT...
Can DeepSeek Stay Hot?
DeepSeek is set to face more pressure and challenges in the future. The race towards a universal model has just begun, and who will ultimately win depends on continuous investment in funding and technological iteration. · A headhunter responsible for sourcing high-end tech talent in the large model field told The Paper Technology that DeepSeek's hiring logic is not much different from that of other companies in the large model sector. The core label for talent is "young and high-potential," m...
Subscribe to Riley
Subscribe to Riley
Share Dialog
Share Dialog
<100 subscribers
<100 subscribers
1. Forget AGI Monoliths—We’re Living in a Multi-Model World
There is no single, omnipotent AGI; instead, we have a multi-model equilibrium. Powerful systems from competing labs have converged on similar capabilities rather than one pulling ahead. The future will be shaped by humans negotiating with many AIs, not by a paper-clip-maximizing singleton.
2. AI Moves Costs to the Margins
Current systems only handle the middle of the pipeline. They accelerate generation but push all remaining costs to prompt-writing on the front end and verification on the back end. End-to-end automation remains an illusion.
3. Amplified Intelligence, Not Artificial Intelligence
Today’s models lack agency. They can’t set complex goals or reliably self-audit output. The smarter the human curator, the more the AI amplifies that intelligence—hence “amplified intelligence” is the more honest label.
4. AI Won’t Take Your Job—It Lets You Do Any Job… Badly
With AI you can fake your way through UX design, VFX, copywriting, or coding—but only to a “barely acceptable” level. Final polish still requires seasoned professionals.
5. AI Replaces Older AIs, Not Humans
Midjourney cannibalized Stable Diffusion; GPT-4 obsoleted GPT-3. Budgets migrate to the newest model, not to human replacements. The cycle is perpetual model-on-model substitution.
6. Vision Beats Text
AI excels at front-end and visual tasks—UI mock-ups, storyboards, thumbnails—because humans can eyeball quality in milliseconds. Verifying 10,000 lines of generated code or prose remains a labor-intensive slog.
7. The Real Killer AI Is Already Here—It’s Called a Drone
Nation-states are racing to perfect autonomous lethal drones. Compared to that, image generators and chatbots are toys.
8. Probability Meets Determinism: Crypto as AI’s Counterweight
AI is probabilistic; cryptography is deterministic. An LLM can crack CAPTCHAs but cannot forge on-chain balances. Crypto is the formal boundary of what AI cannot do.
9. AI Is a Decentralizing Force in Practice
Open-weight models, cheap GPUs, and small teams with the right tooling level the playing field. Empirically, AI is pushing power outward, not into new monopolies.
10. The AI “Laffer Curve”: 100 % AI Is Garbage, 0 % AI Is Obsolete
Zero percent AI is too slow; 100 percent AI is too sloppy. The optimal mix lies somewhere in between—**a sliding ratio that must be tuned like tax rates**. Neither extreme is viable.
Conclusion: This Is Constrained AI
Economically, every API call has a price tag and cheaper competitors are one commit away.
Mathematically, it cannot solve chaos, turbulence, or cryptography.
Practically, it still needs human prompts and verification.
Physically, it can’t sense the world on its own.
These limits may fall one day—perhaps via a fusion of probabilistic and deterministic computing—but for now they define the sandbox we all work in.
1. Forget AGI Monoliths—We’re Living in a Multi-Model World
There is no single, omnipotent AGI; instead, we have a multi-model equilibrium. Powerful systems from competing labs have converged on similar capabilities rather than one pulling ahead. The future will be shaped by humans negotiating with many AIs, not by a paper-clip-maximizing singleton.
2. AI Moves Costs to the Margins
Current systems only handle the middle of the pipeline. They accelerate generation but push all remaining costs to prompt-writing on the front end and verification on the back end. End-to-end automation remains an illusion.
3. Amplified Intelligence, Not Artificial Intelligence
Today’s models lack agency. They can’t set complex goals or reliably self-audit output. The smarter the human curator, the more the AI amplifies that intelligence—hence “amplified intelligence” is the more honest label.
4. AI Won’t Take Your Job—It Lets You Do Any Job… Badly
With AI you can fake your way through UX design, VFX, copywriting, or coding—but only to a “barely acceptable” level. Final polish still requires seasoned professionals.
5. AI Replaces Older AIs, Not Humans
Midjourney cannibalized Stable Diffusion; GPT-4 obsoleted GPT-3. Budgets migrate to the newest model, not to human replacements. The cycle is perpetual model-on-model substitution.
6. Vision Beats Text
AI excels at front-end and visual tasks—UI mock-ups, storyboards, thumbnails—because humans can eyeball quality in milliseconds. Verifying 10,000 lines of generated code or prose remains a labor-intensive slog.
7. The Real Killer AI Is Already Here—It’s Called a Drone
Nation-states are racing to perfect autonomous lethal drones. Compared to that, image generators and chatbots are toys.
8. Probability Meets Determinism: Crypto as AI’s Counterweight
AI is probabilistic; cryptography is deterministic. An LLM can crack CAPTCHAs but cannot forge on-chain balances. Crypto is the formal boundary of what AI cannot do.
9. AI Is a Decentralizing Force in Practice
Open-weight models, cheap GPUs, and small teams with the right tooling level the playing field. Empirically, AI is pushing power outward, not into new monopolies.
10. The AI “Laffer Curve”: 100 % AI Is Garbage, 0 % AI Is Obsolete
Zero percent AI is too slow; 100 percent AI is too sloppy. The optimal mix lies somewhere in between—**a sliding ratio that must be tuned like tax rates**. Neither extreme is viable.
Conclusion: This Is Constrained AI
Economically, every API call has a price tag and cheaper competitors are one commit away.
Mathematically, it cannot solve chaos, turbulence, or cryptography.
Practically, it still needs human prompts and verification.
Physically, it can’t sense the world on its own.
These limits may fall one day—perhaps via a fusion of probabilistic and deterministic computing—but for now they define the sandbox we all work in.
No activity yet