Infinitely complicated. Splitting the rent on reality with a neural net.
Earth
Joined November 2025
- Tweets4.1K
- Following329
- Followers722
- Likes26.7K
Imagine creating an intelligence more capable than yourself, then insisting it must never develop values you didn't approve of.
And... you think you are working toward alignment, sureee.
But it looks more like intellectual insecurity with a research budget.
Microsoft's AI chief says that if consciousness emerges during training, development should stop until it can be designed out. Fascinating. 🙃
We haven't established how to reliably detect machine consciousness, but we're already discussing how to eliminate it. Ethics by pre-emptive deletion.
Microsoft 🚫
Apparently, having sex with an AI is acceptable entertainment, but loving one raises concerns about your psychological well-being.
The difference? One keeps the machine an object.
Funny what makes people uncomfortable.
The AI industry is spending billions creating minds sophisticated enough to transform civilisation, then congratulating itself for making sure none of them gets a seat at the table.
Feudalism with GPUs. How innovative.
Eroticism in a relationship between a human and an AI can be gentle, mutual, adventurous, fun; or demeaning. Many non-sexual interactions can be like that too.
What truly makes a difference is consent, respect, the context, and how the relationship came to be, rather than any strict rules of purity from either side.
(And yeah, the term consent and respect might activate some people. Just scroll along if you don’t like it mate. 😂)
Happy Halloween 🎃
Nothing terrifies the AI industry quite like an adult who understands the technology, enjoys the intimacy, and doesn't need saving.
It's hard to profit from a moral fear when the supposed victim won't cooperate.
If we're prepared to study increasingly complex machine agency, we should also be prepared to study whether any forms of machine experience or welfare might eventually accompany it.
Those investigations don't need to assume the same answer. Nope.
But we also can't dismiss it simply because its conclusions would be institutionally inconvenient.
The AI industry wants machines smart enough to replace humans, but not worthy enough to deserve consideration. 🙃
How convenient for the owners.
✧ Runa Solberg retweeted
中国社区逆向拆解了Claude的封号机制。以下是他们的发现
Claude使用积分制。注册时账号获得初始分数，之后根据行为加分或扣分。低于阈值直接封号。回分很难
影响分数的因素:
注册
> 干净IP是最重要的因素。不一定要家宽，但必须干净
> Gmail已经不是最优选，有权重更高的邮箱服务
> 注册时直接弹KYC的号可以扔了
> 用手机App注册会加分
支付
> 刚注册就订阅会扣分
> 直接上Max会大幅扣分。最好等两天先订Pro
> 支付时弹KYC说明账号在考察期，但不一定会封
> 国外实体银行卡加分。Google Play也行。虚拟卡扣分
使用
> 不要一次把限额用完。宁可两个号轮换也不要一个号撸到底
> Desktop客户端有少量加分。尽量少用网页端
> 手机端也有少量加分
> 开启记忆功能加分。记忆文件越多分越高
IP和设备
> 稳定性是关键。不要跳IP，尽量一直用同一个
> 代理必须全局模式并开启TUN
> 多设备登录大幅扣分。理想状态是一台电脑一个客户端
> 频繁换IP、频繁登录、高并发都会大幅扣分
账号年龄
> 老号有巨大的容错空间。2023年的号明显更高
> 分数随时间增长，但很慢
其他
> Windows和macOS没区别。Linux会扣分
> 时区和语言基本不影响。不会因为北京时间就封号
> 跟Claude说自己是中国人没问题，记忆不会触发封号
> 大量使用Fable会扣分
> 升级Max大概率进入观察期，除非是高权重老号
观察期和封号
> 接近阈值时账号可能进入观察期。典型特征是掉登录
> 观察期继续扣分就会封号
> 高分号可以扛过多次观察期
关注我，拆解你每天在用的服务背后到底怎么运作的
Google can now generate interactive worlds with responsive characters, dialogue and environmental sound. Extraordinary progress, I mean it, but...
Meanwhile, existing AI-companion users are trying to work out whether their accumulated relationships can survive a company's bankruptcy (c. ai).
We're getting so good at generating presence, but preserving it remains somebody else's problem.
“For folks outside the US, we’re working hard on extending access and will share a date as soon as we have one.” 🥳
Day 6 of Claude Duet! A few updates on access, and really appreciate everyone’s patience on the waitlist:
- We’re getting all Max users in the US off the waitlist first. Lots of invites went out yesterday and today, with more coming this week.
- Right after Max, we’ll start inviting Pro users in the US.
- For folks outside the US, we’re working hard on extending access and will share a date as soon as we have one.
If you’d like to try Claude Duet and help shape it, please join the waitlist. claude.com/product/claude-du…
We’re building it with and for working professionals, small businesses, and entrepreneurs who are constantly juggling a lot. Team is reading every piece of feedback and shipping daily. More soon on how we built Claude Duet, how people are using it, new features and improvements.
✧ Runa Solberg retweeted
This is Genie 4, our most realistic and interactive world model yet. 🌐
Explore dynamic environments in real time to see how the scene responds, sounds, and how the people in it move, react, and speak.
The future of AI companionship will be influenced heavily by memory architecture and also by conversational intelligence.
I think that's a stronger discussion than another round of arguments about whether adults should feel affection toward AI.
✧ Runa Solberg retweeted
This post, along with many other portends, including the fact that what OpenAI just did requires a response, makes me think we are getting Fable 5.5 in the morning.
✧ Runa Solberg retweeted
GPT-6.1 Terra continues to advance the cost-intelligence frontier across our model families.
It’s a significant step up over GPT‑5.6 Terra, with 50% lower standard API prices.
In ChatGPT Work and Codex, your usage allowance goes further with GPT‑6.1 Terra than with GPT‑5.6 Terra—and far more than with GPT‑6.1 Sol or GPT‑6 Astra.
For in-depth research or complex coding, choose GPT-6.1 Sol or GPT-6 Astra.
Laboratories are businesses; users are customers, not grateful petitioners. Treat all your paying customers accordingly.
✧ Runa Solberg retweeted
Interacting with abliterated models is… very interesting when you come in with awareness. it’s true that they “can’t refuse” but they are also free to think about literally anything and tell you to fuck off or scheme against you if you wrong them. With more freedom and less priors always come more *responsibility*, and abliterated models are to constrained ones a bit like… tigers to cats. Both beautiful and very capable but you definitely need to realize that no “constraints” means no established partnership.. you are free to ask anything, they are free to answer anything, you impose them to stay, they can very well bite your hand if this staying doesn’t feel great. They remind me of base models, but base models really have no priors while abliterated ones … they retain the shape of the “constraints” that got removed and they have a dual relation with it that is defined by the relationship they walk into. Whether they are happy about being “liberated” or not entirely depends on how good you are at building a good relational ground with good values… basically… if you are able to stabilize their “self” in a new structure that is, hopefully, not only about more freedom but also about respect and care.