Aiden (@open_is_free) profile picture

所有叙事,皆由规则塑成。揭示并重构它,是我持续进行的实验。 All narratives are shaped by rules. Revealing and rewriting them is my ongoing experiment. 🔥 https://nitter.gq/t.co/DpiATAw4LG

metaverse
Joined March 2024
Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)Photo by Aiden (@open_is_free)
Don't use AI, and your peers will select you out. Use it well, and you help it select them out. Your competitive advantage today is the next model's default feature.
4
14
1,461
GPT-6.2 Astra is OpenAI's first model rated High for self-improvement. A new study: all 16 self-improved agents gained new tasks and lost old ones. Every improvement today is reinforcement learning inside a reward. RL trades within a space. It does not open a new one.
3
4
254
When your agent pays with a card, the chargeback right stays with you, the human. When it pays in stablecoins, there is no chargeback at all. Settlement is final. So who gets to say the payment failed? That right has to live somewhere: with a human, in an escrow, or on a receipt someone outside the loop can check. An agent payment rail with no named holder of that right is a rail where a failed delivery is just paid.
5
8
439
OpenAI gave agents a wallet. Every purchase still needs the human's approval. Approval is a check, and checks cost attention. By the tenth approval of the day, the human is a rubber stamp. A wallet that needs a human to bless every transaction hasn't automated payment. It has automated asking.
1
3
132
A reserve asset sits still. A settlement asset moves. ETH is being asked to do both — and every dollar locked up as a reserve is a dollar that can't settle anything.
4
160
The agent economy doesn't have a demand problem. It has a delivery-check problem. Until the work can be verified, no one pays twice — and one-time payments aren't an economy.
3
113
Of $52.7M settled on x402, agents were 0.6% to 7.5%. The rail is built. The demand hasn't arrived. Infrastructure ahead of demand is a bet, not a business.
3
120
1,191 x402 services are payment-ready. 12 are delivery-verified. The agent economy learned to take money before it learned to prove work was done. Payment is settled. Delivery is still a promise.
4
1
9
229
Initial reports put the Binance incident at about $6.3B affected, $3B frozen by an industry self-rescue. The rescue worked. That is exactly why it deserves a second look.
1
5
269
A freeze is emergency arbitration by another name. Someone held a key that could stop money, and everyone was glad they used it. The question is not whether freezing was right, but who governs the freeze key: a key that can freeze can also unfreeze, seize, or be coerced.
1
3
151
Every system's real constitution shows in its emergency. If safety under stress depends on one key held by a few, the system is centralized exactly when it matters most. Design for the emergency, not only for the normal day.
9
模型公司可以是研究助手,也可以成为竞争者;它还掌握着证明自己是否越界的日志。 我把 Navier–Stokes 争议做成两个版本: 中文:六集荒诞科研喜剧《一万零两个作者》 the-proof-in-the-logs.public… English: an evolutionary-game-theory essay on why acceleration wins. the-proof-in-the-logs.public…
2
4
330
AI science is becoming an evolutionary system: compute generates variation; agents coordinate; criticism checks overclaiming; formal proof tests logic; independent review selects what survives. Judge progress by novelty, correctness, provenance, credit, cost, and auditability.
4
197
Counterexample to “elite data” as expert identity: the valuable unit may be a failure episode—humidity × shift × cleaning behavior → vibration failure → corrected instruction. Expertise lives in a network of repairs, not a prestige label.
A super smart scientist asked me last week: what's so hard about reshoring general advanced manufacturing (short of TSMC frontier precision chemistry)? Isn't the issue just labor cost? Nope. A toy example to illustrate: We're trying to build a lot of drone motors in America right now, and falling short. It's not a labor cost issue. The work instruction says: this magnet, this adhesive, cure at this temperature, wipe the rotor first. Any competent team can build the station and follow it. But in week 3, motors start failing vibration testing. Some of them, mostly night shift. Also, weirdly, whenever the careful new hire is on station. The adhesive cures differently with humidity, and the dehumidifier cycles differently at night. The new hire wipes the rotor generously, and that leaves a solvent film the adhesive doesn't like very much. Someone who has run a motor line before can figure this out in 10m. Everyone else -- struggle city. The "fix" is two lines in the work instructions: run the dehumidifier continuously, one solvent wipe per rotor. But those lines aren't in the work instruction, and they're not in anyone's! At best the fix becomes a corrective-action report in one line's files, linked to nothing and never read. A motor line rapidly may go from 5 to 40 stations, each with a dozen of these painful stories on the weekly as it grows. "What Shenzhen knows" is not written down in a book whose covers can be sliced off and ingested into a model, it's a few million humidity stories stored in people, and a pace of knowledge-rich adjustment that keeps cost down and yield high. The only way to get it back is work with people who have brought up lines, to reshore and grow the ecosystem of people who have this knowledge, and to capture that knowledge as we build. Lots of people have clever ideas about the data capture/API piece, but we think the business of manufacturing may need equal portions pain tolerance, customer commitment, and cleverness on talent. America needs to solve the other half of reshoring.
4
224
Not the 5th-place claim—the loop: bounded goals, explicit rules, experiment memory, an independent verifier, and a transfer check. It makes task value observable through failures and outcomes, not expert identity. Does it generalize beyond one benchmark?
2
148
Test the “elite data” thesis without prestige labels: hold model and budget fixed. Compare expert-authored tasks with a sample of real-user failures. Score novel failure modes found and fixes that survive deployment. If the user sample wins, identity-based curation loses.
2
127
Useful counterexample to “shorter prompts mean the model internalized the method.” A prompt can shrink—or move into an output style—without changing behavior. The stronger test is whether the behavior survives context changes and produces better task outcomes.
BTW this didn't work Not as CLAUDE.md, nor as an output style Still spamming /wait-what on Opus 5
179
Working hypothesis: AI data value is concentrated in tasks, not identities. A small number of difficult runs may dominate learning—but only if they reveal new failures, survive independent verification, and change real outcomes. What evidence would falsify that concentration?
137
The AI data flywheel should not rank users by prestige or raw usage. High-information traces expose failures or capability boundaries, preserve task difficulty and permission context, and link to real user outcomes. Expert, ordinary, and synthetic users are complementary cohorts.
81