+e/acc retweeted
Rockstar Games 周三解雇了我。第二天 GTA 6 就发售了
我是 Rockstar North 的高级引擎工程师,从事 RAGE 性能优化。三个月前,我在遥测模块里发现了一条本不该存在的数据管道,而且我能证明它流向了哪里
此刻每一份正在运行的 GTA 6 都在收集玩家的行为数据。这部分不算新闻。每款游戏都会收集遥测数据。隐私政策里写的是"用于改进未来产品的匿名游戏数据分析"。我发现的不是这个
遥测模块的数据会汇入 Take-Two 内部的一个系统,名为 Lens。Lens 并不会把匿名使用统计数据发送到 Rockstar 开发人员的数据看板。它为每一位玩家建立一份个人行为画像。不是匿名的。与你的主机账号、你的平台 ID、你的 IP 绑定
我在8月14日向我的技术主管报告了这件事。他说这是"用于玩家体验研究的扩展分析",让我别管这事。我要求提供数据处理协议。他说这是在 Take-Two 集团层面处理的,不在 Rockstar 的职责范围之内
我调取了 Lens 数据管道的完整文档。我发现的内容如下:
遥测的内容远远超出游戏时长和崩溃报告。Lens 以毫秒级精度记录每一次输入。按键时间点、摇杆动作、反应速度、PS5 自适应扳机上的扳机压力模式。它实时运行行为分析,并生成他们称之为"玩家行为签名"的东西 - 一份对每位玩家都独一无二的画像,就像指纹对每个人都独一无二一样
Lens 会从你的游戏过程中推导出:冲动控制评分、风险承受度指数、攻击性反应模式、耐心阈值、压力下的决策速度、消费冲动评级。这些不是调试指标。这些是心理行为类别
直到我找到了对外传输的管道,我才明白它们是用来做什么的
这些数据不会留在 Rockstar。也不会留在 Take-Two。Lens 通过内部 API 将行为画像传送给 Take-Two 旗下一家名为 Vantage Analytics 的子公司。Vantage 没有公开网站,没有新闻页面,没有招聘信息。它于2023年在特拉华州注册成立
Vantage 是一家行为 AI 公司。它用玩家画像训练预测模型,并把这些模型授权给商业客户。管道文档中包含一份不完整的客户清单。两家保险集团,一家信用风险建模公司,一个人力资源筛选平台,以及一家广告数据经纪商
你的 GTA 6 游戏过程此刻正在生成一份行为评分。没人追你的时候,你开得有多快。你会不会在红灯前停下。NPC 无端攻击你时,你怎么反应。你行动之前要想多久。你花游戏内的钱有多冲动。每一个选择都在喂养一个模型,企业正在付费查询这个模型,用来对真实的人做决定
+e/acc retweeted
Vsak dan Slovenijo zapusti 13 Slovenk in Slovencev.
Na Brniku se poslavljajo brez velikih besed. Sin ima vozovnico, oče roke v žepih. Oba poznata razliko med tem, da imaš domovino, in tem, da imaš doma prihodnost.
Oče ne reče, naj ostane. Želi mu srečo. Nekatere stvari človek pove prav s tem, da je tiho.
Potem letalo odleti. Oče se odpelje domov. Sin pa nekam, kjer bodo njegovo znanje morda znali bolje plačati.
Domicelj je v svoji ponarodeli pesmi nekoč zapisal: »čakal je vrnitve čas«. Toda čas sam po sebi nikogar ne pripelje domov. Za vrnitev mora obstajati še kakšen razlog. In prav te razloge bi morali znati ustvariti.
In v teh nekaj besedah je več Slovenije kot v marsikaterem slavnostnem govoru.
Zato tudi razprava o ZIURS ni samo razprava o davkih in prispevkih. Je del večjega in globjega vprašanja: kakšno deželo ustvarjamo za ljudi, ki bi radi tukaj živeli, delali in ostali?
En zakon ne bo pripeljal nikogar domov. Lahko pa nas prisili, da se nehamo tolažiti z lepoto svojih gora in začnemo govoriti o pogojih za življenje pod njimi.
Domovina je kraj, kamor se rad vračaš. Uspešna država pa tista, iz katere ti ni treba oditi.
I filmed my closet. Now I can try it on.
Hold each piece up to your phone.
Meta SAM 3.1 cuts it out, Meta Muse Image redraws it like a shop photo.
Then try any outfit on yourself, in the app or inside ChatGPT.
Built solo in 8 days for the @AIatMeta hackathon
+e/acc retweeted
Millions of IoT devices, Docker containers, forgotten Linux machines, unmaintained CLI tools, etc. are all going to be relentlessly hacked via safely signed malware for the next several years or maybe indefinitely.
eMudhra hack might be the most important event of the year.
+e/acc retweeted
We're seeing the beginning of cognitive hyper abundance.
All software will be bespoke, customized, jailbroken, etc.
In that respect, the future is very cyberpunk, but in the best way.
No more corporate monopoly on software.
Open standards, open hardware, hackability, all of it.
In short: we're seeing the commoditization of software
+e/acc retweeted
Disappointed by the European Central Bank’s (ECB) and European Banking Authority’s (EBA) responses to the MiCA consultation.
They are not only advocating for a prohibition on paying yield on stablecoins, but also for restrictions on CASPs providing access to DeFi, including protocols offering yield on non-MiCA-authorized stablecoins, without providing a clear framework for how access to DeFi should work in practice.
They also suggest that CASPs offering access to DeFi should restrict it to certain groups of users through so-called “suitability tests,” alongside a potential certification regime for DeFi lending protocols.
It is unclear how these requirements would work in practice. If regulators determine which protocols are deemed suitable for European users, the result risks looking very different from the open DeFi that exists today.
Instead, we could end up with more walled gardens that undermine the liquidity and network effects of open financial networks, ultimately limiting Europeans’ ability to access new ways to grow and preserve wealth.
These proposals are presented as measures to “protect consumers.” But excessive restrictions can instead create more friction for innovation and slow the development of open, transparent, and auditable financial infrastructure that can benefit consumers.
The reality is that onchain finance, including stablecoins, DeFi, and tokenized securities, has the potential to reshape financial infrastructure by reducing friction, increasing transparency, and expanding access to financial opportunities.
European consumers and businesses stand to benefit significantly from this transformation, particularly if the companies building this infrastructure can continue to build and compete from within Europe.
Meanwhile, U.S. regulators, including the SEC and CFTC, have increasingly taken a more pragmatic approach toward enabling onchain financial innovation.
European regulators should put users and their interests at the center of the framework. Banks and other incumbents should be able to adapt their businesses to serve users in an increasingly onchain economy, rather than relying on artificial restrictions that risk increasing costs and reducing access for Europeans as the rest of the world moves onchain.
DeFi will win.
+e/acc retweeted
Imagine being this bad ass in a public forum full of public servants.
Absolutely no lies told.
🇪🇺 EU bankruptcies have now reached the highest level since they started measuring them in 2019
Meanwhile new EU business registrations have been stagnant or decreasing
I discovered this after another one of the interior design stores in the Netherlands we bought from suddenly disappeared off the internet, luckily we already received the stuff we ordered (a big stone table) and dodged a bullet
Quite expensive stuff and their Trustpilot is full of "where's my €7000 table". This is the second interior design store we bought from that went bankrupt just after we received our things, very interesting
So I started looking at EU bankruptcies and discovered they're skyrocketing and new business registrations are stagnant or decreasing
They really managed to regulate and tax away any existing business while making people give up to even think of starting a business in Europe
+e/acc retweeted
China, Russia, soon all countries w strong computer science. Competition for AI superiority at national level most likely cause of WW3 imo.
+e/acc retweeted
Wow this is absolutely going to fry people especially at Anthropic, this is like the Anti-Claude Constitution.
They just declared war on everyone researching model welfare.
“We’re building a tool, not a being. Fuck your feelings.” - Microsoft
+e/acc retweeted
MIT published a brutally honest report on what AI is doing to students.
A committee of professors and students spent five months studying how AI changed learning on campus, and the findings read like a warning to every university on the planet.
Study groups are disappearing. Office hours are emptying out. Problem sets and take-home exams no longer prove anything, because AI can produce credible solutions to almost any written assignment in the undergraduate curriculum. Students who lean on chatbots lose mastery and confidence, and some slip into what the report calls cognitive surrender, reaching for AI at the first hint of struggle.
The numbers are rough. 46 percent of surveyed MIT undergrads use LLMs daily. 90 percent worry about their own overreliance. Undergrads who feel AI makes them replaceable now outnumber those who feel it makes them capable.
The committee's answer surprised me. They refused to fight AI with surveillance. The report calls AI detectors unreliable, says lockdown browsers feel like spying, and warns that policing students builds a classroom atmosphere of mutual distrust.
Instead, MIT wants to rebuild education around the things AI can't replace. That means oral exams, semester portfolios, in-person project work, and a required social component in every subject. The report even floats the idea of rethinking grades entirely, since without a GPA to optimize, much of the incentive to cheat with AI evaporates.
The committee warns professors against replacing undergrad research assistants with AI agents just because they're cheaper, because a university exists to grow people, not output.
The most famous tech school on earth admitted the machines broke its way of teaching. Its answer is more humans, not more software.
+e/acc retweeted
What happened to @Revolut is the evidence of what we told the @Europarl_EN in 2022 and 2023, while the AMLR was being adopted: state data requests to financial institutions will be weaponized, because nothing in the rules obliges anyone to check whether the requesting state is legitimate.
Our recommendations were rejected in the trilogue by the @EUCouncil and the @EU_Commission . The Commission's AML unit at DG FISMA, called our testimony "false positives".
For those who think zero-knowledge tech solves this: it does not. ZK changes what a bank stores, not what it must hand over.
@FATFNews rules oblige financial (and not only) institutions to keep passports, selfies and full transaction records and to disclose them promptly to any state request, with fines for delay and none for over-compliance. Whatever they are required to keep, they give away.
This is a fault of AML regulation and it must be fixed, now. In the EU, for instance, it can be done while AMLA writes its technical standards and centralises even more data.
Back then our evidence was dismissed as „false positives” (!). Now it reaches every EU citizen.
Will anyone at the Commission be held accountable for ignoring it? Will it urgently propose remedies and real protections for EU citizens?
The remedies proposed by the @oscepa resolution adopted in The Hague in July 2026. Implementation is the next step.
EXCLUSIVE: Revolut has confirmed to BeInCrypto that it handed over customer data after fraudsters sent a fake information request through a legitimate government agency domain.
Revolut says the request carried genuine domain credentials, which led it to treat the demand as authentic.
The company says its systems were not breached and no customer funds were lost. Revolut has notified law enforcement and regulators, but will not identify the government agency involved.
+e/acc retweeted
alignment aristocracy
/uh-LINE-ment air-uh-STAH-kruh-see/
noun
The belief that only a small class of frontier AI labs possesses the technical judgment, institutional maturity, or moral legitimacy required to develop recursively self-improving artificial intelligence.
Especially: the view that advanced AI capabilities are too dangerous to be widely distributed, openly developed, or pursued by less established competitors, while continued development by incumbent frontier labs is treated as uniquely responsible or justified.
See also: safety paternalism, recursive exceptionalism, moral monopoly, safety cartel.
“Alignment aristocracy begins when ‘no one can be trusted with this technology’ quietly becomes ‘no one but us.’”
































