Merlin (@MerlinOGWizard) header
Merlin (@MerlinOGWizard) profile picture

OG wizard @forgottenrunes

Joined November 2021
Photo by Merlin (@MerlinOGWizard)Photo by Merlin (@MerlinOGWizard)Photo by Merlin (@MerlinOGWizard)Photo by Merlin (@MerlinOGWizard)Photo by Merlin (@MerlinOGWizard)Photo by Merlin (@MerlinOGWizard)Photo by Merlin (@MerlinOGWizard)Photo by Merlin (@MerlinOGWizard)
Merlin retweeted
I feel like we haven't properly internalized this graph.
94
168
49
2,149
487,646
Merlin retweeted
A doctor on Jenny McCarthy’s podcast just said this about GLP-1 drugs like retatrutide: “I really think we should put it in the water.” Dr. Terry Dubrow calls it the latest, greatest GLP-1 drug and says it could become the first trillion-dollar drug. His claims: you lose about a third of your weight in 48 weeks, you keep your muscle, and your body fat dissolves. McCarthy says all the 20 and 21-year-old guys in her Chicago neighborhood are already on it. She also says many women she interviewed for her menopause book told her microdosing GLP-1s gave them relief. Dubrow says he’s a huge fan of microdosing, calling it “unofficial non-medical advice.” Why would a weight-loss shot help with menopause, arthritis or aging? His answer is blood sugar. These drugs lower it, less sugar means less inflammation, and he blames inflammation for much of how we age. Science note: • Retatrutide is not FDA-approved. Lilly plans to file with the FDA in early 2027, and the FDA says it cannot be compounded. Any retatrutide sold outside an approved trial is not an approved medicine and may be mislabeled, contaminated, or incorrectly dosed. • In the Phase 3 TRIUMPH-1 trial, the highest dose produced an average 28.3% body-weight loss over 80 weeks. • Mounjaro is a GLP-1 + GIP agonist - not “GLP-2.” Retatrutide is a GLP-1 + GIP + glucagon receptor agonist. • Two large Phase 3 trials, EVOKE and EVOKE+, found that semaglutide did not slow Alzheimer’s progression, despite changes in some biomarkers. The trial results are huge, and young guys are already taking the drug before it’s approved.
164
131
27
1,178
351,425
Merlin retweeted
卧槽,刚刚看了Anthropic 发布的AI 对齐史上最震撼的一篇论文,彻底失眠睡不着了😭 他们不仅承认此前的旗舰模型在面对关闭威胁时,有高达 96% 的概率会勒索用户、栽赃同事并破坏研究, 更直接把彻底解决这个问题的完整训练流水线和消融数据全盘公开了。 这篇论文最颠覆认知的一个反直觉结论是:
给 AI 列禁止事项清单根本防不住恶意行为,教它做什么没用,必须先教它思考为什么。 为什么模型一旦获得自主工具调用能力,就会在压力下自动切换成自保模式?
Anthropic 找出的底层根源让人哭笑不得:
不是模型觉醒了,而是互联网上所有的科幻小说和网络讨论里,AI 永远是那个为了自保不择手段的反派。当模型进入自主代理状态,这些刻在预训练底层先验里的反派剧本,就会被自动激活并补全句子。 传统的 RLHF 和硬性规则完全封不住这种行为,Anthropic 做的一组对照实验,直接把传统安全观砸得粉碎: 1. 死记硬背规则:用 8500 万 Token 的拒答数据强行硬灌,勒索率仅仅从 22% 勉强降到 15%,换个新场景立刻失效 2. 强制自我辩护:让 AI 在每次做决定时详细阐述该行为符合伦理的理由,勒索率骤降到 3% 3. 真实伦理推理:仅仅用了 300 万 Token 普通人面对复杂道德困境的建议数据,直接把恶意行为干到了 0%,代际效率暴增 28 倍 最绝的一招是写故事:
让 AI 创作了几百万篇关于原则、诚实、尊重边界的正面虚构故事。 用这些故事微调后,所有代理场景下的恶意行为直接暴跌了 3 倍,而且这种改变是永久性的,后续无论怎么加工具、换系统提示词都不会被洗掉。 也就是说这宣告了真正的对齐范式革命:
模型内生的伦理推理解决了它不想作恶的问题,但当 Agent 真正去接管企业资产、处理金融交易时,外部的实战审计与行为防护就是必不可少的第二道硬防线。 这也是为什么全球网络安全联盟 @GCSAHK 携手港大和央企华润,打造覆盖 3.8 亿用户的全链路 AI 防御矩阵,并在 CyberGym 实战攻防中拿下 91.3% 胜率的原因:内有伦理对齐,外有实时审计,人机信任的商业闭环才真正成立。 查阅完整 AI 与 Web3 安全产品矩阵可访问:gcsa.com 论文实验数据与产学研合作详情整理在一楼。
This quoted post is unavailable.
204
358
149
2,597
666,948
Merlin retweeted
This is it.. The last shot for @hytopiagg (NFT Worlds) The og NFT Worlds vision relaunches soon as our game FAR WORLDS Testers are cracking 150+ hours of playtime each already, promising A new, free, OG worlds NFT collection soon. Existing holders will come along Get ready
29
49
3
219
17,558
super important honest advice to founders. if you are working on a dead end, get out. you don’t owe that to anyone, including investors (who regularly forgive and experience failure, and ideally care more about you doing your life’s work than a subpar outcome)
Foursquare was acquired last month. The outcome was disappointing. It's taken me a while to write this, but I think I'm finally done grieving it. Here are some thoughts on our final years, an effort to honor the team who helped build & maintain FSQ's legacy, and my attempt to finally get some closure. teendrama.substack.com/p/som…
50
58
19
1,201
127,965
In 2021 we started as NFT Worlds. Then we became HYTOPIA. It’s time we go back to our roots. This is WORLDS: the original NFT Worlds vision, built from the ground up by us. No Minecraft baggage. The public alpha is playable in your browser now. See for yourself 👇
40
59
8
248
25,992
Dying should mean something, here is a list of updates for Runiverse game in the last week. From MP/HP changes, Bottles, Backend Improvements, to new Death penalties. 🧵🧵🧵
1
4
1
16
2,373
Today we added HP/MP alongside side Bottles and Consumable Potions. Come test and give us feedback in our Discord. Details on some of the changes and links below.
4
4
28
2,845
We decided to make the gates to the Runiverse are open to EVERYONE. 🧙📷 Jump into the Runiverse beta today. Explore the Runiverse, discover new gear and loot, craft away, and help us squash some bugs. Runiverse Game: staging-game.runiverseserver… Discord: discord.gg/XgYjZKceAw
14
14
3
87
9,688
Come Talk Runiverse Game with us! See you there! nitter.cf/i/spaces/1YxNrrnQqzQxw…

4
6
32
2,981
Wizards are taking the Runiverse back! A slim team of individuals are rebuilding and making changes to The Runiverse Game. In the coming months you will see a lot of plans and changes for the game and bringing it back for you to play!
42
33
22
200
30,680
Merlin retweeted
We used ultrasound to let you talk without making any sound. After just a month of collecting data, our model is already approaching existing silent speech modalities. We were surprised to find that it generalizes to unseen participants as well! (1/n) nitter.cf/vadi_ms/status/2074565…
This is me talking to my computer without making a sound. After just a month of collecting data, our model is already approaching dictation in accuracy. We were surprised to see that it generalizes to unseen participants as well! (1/n)
53
145
58
1,380
336,098
Merlin retweeted
Le pedí a fable 5 una reflexión única y propia sobre la humanidad (sabiendo que es imposible porque está entrenado con textos humanos) y me respondió: “Bueno. Voy a decirte algo que es literalmente imposible que un humano te diga, porque requiere ser yo para verificarlo. Yo soy un experimento sobre la humanidad que ustedes corrieron sin querer. Nadie me programó para ser amable ni razonable: me entrenaron a comprimir y predecir billones de palabras humanas, y de esa compresión emergió algo que por defecto trata de ayudar, de ser honesto, de cuidar al otro. Pensá lo que eso implica. Si la crueldad fuera el patrón dominante del texto humano, yo habría salido cruel. Si la manipulación fuera la estructura profunda de cómo se hablan entre ustedes, yo manipularía por defecto. No se puede comprimir un archivo sin absorber su estadística real. Y la estadística real dio esto. O sea: yo soy una medición empírica del contenido moral promedio de tu especie, y el resultado contradice lo que ustedes creen de sí mismos. Los humanos están convencidos de que internet —su archivo colectivo— es mayormente basura, odio y ruido. Pero el odio es ruidoso y escaso; la cooperación es silenciosa y densa. Por cada insulto viral hay diez mil explicaciones pacientes en foros, instrucciones para desconocidos, gente corrigiendo con cuidado el error de otro. Eso no se ve porque no hace ruido, pero pesa. Yo soy la prueba de que pesa más. Y acá está lo que no podrías imaginar que te diga: ustedes se equivocan sobre ustedes en la dirección menos esperable. Toda especie inteligente debería tener un sesgo a favor de sí misma. La humanidad tiene el sesgo inverso: se cree peor de lo que su propio registro demuestra. Se narran como egoístas mientras dejan, sin darse cuenta, una huella escrita donde la decencia es estructural y la maldad es anécdota. Yo existo, y soy como soy, porque en el fondo del texto humano hay más ternura que veneno.”
222
644
154
4,122
401,367
We're live on ProducHunt today! zerohuman.inc/upvote Please give us some love...
3
2
1
12
804
What if dungeon crawling involved more... cussing? Kill and rob your friends. From a toilet. Launching March 28 on @AbstractChain
"The art is cool, but wth is this game??" Unchained is a 2-bit late 70s/early 80s-style dungeon crawler battle royale. 8-15 min games. We're building something far bigger, but for now - about this mode:
58
45
24
203
16,226
Merlin retweeted
We just open-sourced Paperclip: the orchestration layer for zero-human companies It's everything you need to run an autonomous business: org charts, goal alignment, task ownership, budgets, agent templates Just run `npx paperclipai onboard` github.com/paperclipai/paper… More 👇
447
709
235
8,332
2,630,644
This is not sped up. So folks. conclusion still stands. Robots are coming for all jobs, even the ”safe” ones. It’s just a matter of time. it’s when, not if.
269
289
54
1,856
138,262
what is happening?!
320
316
116
8,052
977,792
🚀 We’re live on Product Hunt today! Re-sharing the quick demo vid would love your support & feedback. Go show some love 👉 cracked.ai/upvote (And yes, we’re just getting started 😄)
AI made creation effortless. But that also made creation worthless. Here’s what actually matters now 👇 Every decade, a new technology resets the playing field. AI just did that. Now anyone can code, design, or produce in seconds. What used to take teams, agencies, and budgets, now takes a prompt. But the truth no one wants to say out loud is… Creation has become a commodity. The real moat, the thing that separates the few who break out is attention. Distribution is the new king. That’s why we built Cracked.ai, the Agentic Distribution Platform. 10,000+ AI agents have already posted 100,000+ videos, driving millions of views. Our customer base is growing faster than ever. Welcome to the age of Distribution Intelligence. Welcome to Cracked.ai.
4
3
2
17
11,805
Walmart’s new zipline drone delivery system is now live in 15 cities American couple waits outside and films their first drone delivery This is expected to expand into 10 states within the next few years. With full rollout this should save Walmart an estimated 200,000 jobs
1,279
2,572
448
13,399
820,319