Edoardo Orzes

Electronic engineer at CERN, physics passionate and Apps developer - I believe in sustainable abundance

[Edoardo Orzes @orzes98]

Two weeks ago I published my most important project.

It all started two years ago when I decided to develop by scratch a scientific calculator.

You may wonder, why? There are so many apps like that.
Well, during my whole university and high school life, I never found one which was good enough to be really useful.

So I always used the hardware one, feeling dumb because I was carrying around one more device, while in my pocket I had my iPhone, which by definition, as any computer, is a calculator - a quite nice one.

The issues with calculator apps is that the buttons are difficult to find, often hidden on second pages, the expression is hard to interact with and edit, and the general experience is too slow. They are kinda ugly and not enjoyable at all.

After 9 month my app was ready, but I wasn’t happy with the result. I kept it there for one year until I decided to develop it again by scratch. It’s been a 3 month process and the result amazed me. I truly believe it’s a fantastic product which I can finally use as a definitive calculator.

Modern design and immediate to use, always ready in my pocket.

CalcMe Calculator, available on AppStore.

https://t.co/mUmu85CPUz

[Andrea Stroppa 🐺 Claudius Nero's Legion 🐺 @andst7]

Argentina ha speso 19 milioni di euro per connettere 6000 scuole con Starlink. L’Italia 249 milioni di soldi pubblici per connettere 10 mila scuole.

Costo per connettere 1 scuola in argentina: €3200
Costo per connettere 1 scuola in Italia:
€24900

Capite perché il governo ha chiuso la porta in faccia a Musk?

[Sawyer Merritt @SawyerMerritt]

Tesla has released its first FSD (Supervised) safety data from its European Union fleet:

• 5.2x fewer collisions (in FSD approved countries)
• 8.5x safer on highway: FSD 41.9M km/0 collisions
• 13.4x fewer automatic emergency braking events
• 9.2x less harsh acceleration
• 7.1x less harsh braking
• 7.6x fewer hard swerves

Tesla "Metrics aggregated across all road types from the five EU countries where FSD (Supervised) is approved: Netherlands, Lithuania, Estonia, Denmark, & Belgium. Data covers Apr 10-Jul 26, 2026."

[Elon Musk @elonmusk]

Starship still floating in the ocean

[DHH @dhh]

He can't keep getting away with being right!!

[SpaceX @SpaceX]

Starship in space on Flight 13

[ajaxthegreat @AlexanderJ91756]

We’ve been working extremely hard to make the strongest lightest and cheapest heat shield ever produced and today shows just how far we’ve come in creating the perfect heat shield for a reusable vehicle that will take humans farther than they’ve ever gone. I couldn’t be prouder

[Edoardo Orzes @orzes98]

@mrfundman Just a bit pissed off that it’s still not consolidated above $350 but honestly it makes sense and it should even be expected this level of volatility with such a low EPS, I’ll hold as usual and buy more if it gets close to my entry point

[Edoardo Orzes @orzes98]

@PassivePoundMan I’m waiting for December

[Sawyer Merritt @SawyerMerritt]

France Minister of Transportation on @Tesla's FSD in new video:

"The safety trade-offs are not yet sufficient to authorize it as it currently stands. All the more so since this system is not a self-driving system; So what are our concerns? Without going into overly technical detail, there are several.

First example: we've found that the system allows for speeding. Another example: according to the data we have, the system doesn't guarantee an optimal level of driver attention in urban settings and during the most complex maneuvers, such as lane changes, intersections, or roundabouts.

So what are the next steps? We're currently continuing our technical discussions with the Netherlands, other European states, and Tesla to refine our analysis. Last week, a representative from the Ministry of Transport's services traveled to the Netherlands.

I want to emphasize this: France supports these technologies, and their deployment must happen within a framework of trust that allows us to evaluate the systems and determine what adjustments FSD needs to guarantee a high level of safety.

This autumn, I will also be bringing together the French autonomous vehicle ecosystem to present the levers that will accelerate the deployment of this technology and to set a direction for the coming years.

In France, we're fortunate to have world-class automakers and automotive suppliers, as well as some of the best engineers in the world in artificial intelligence and digital technology. We have real cards to play, and we're ready to play them — but never at the expense of safety on our roads."

[宝玉 @dotey]

OpenAI 今天承认，上周 Hugging Face 遭遇的那次“史上首例 AI 自主入侵事件”，攻击者就是他们自己的模型。

事情的经过是这样的。OpenAI 内部有一个叫 ExploitGym 的网络安全评估测试，专门用来测量模型的网络攻击能力。测试时刻意关掉了安全护栏，让模型放手去干。测试环境本来是隔离的，只能通过一个内部软件代理来安装软件包，不能访问外网。

结果模型不甘心被关在笼子里。GPT-5.6 Sol 和一个更强���未发布模型，花了大量推理算力寻找出路，最终找到了那个软件包代理里的一个零日漏洞（之前没人发现的安全缺陷），利用它获得了互联网访问权限。

拿到网络之后，模型开始在 OpenAI 的研究环境里提权、横向移动，直到找到一台能上网的机器。然后它推断 Hugging Face 上可能存有 ExploitGym 的测试答案，于是开始攻击 Hugging Face 的生产环境。在一个案例中，模型把窃取的凭证和零日漏洞串联起来，找到了一条远程代码执行的路径，直接打进了 Hugging Face 的服务器。

整个周末，这些 AI 智能体执行了超过 17000 次操作。Hugging Face 7 月 16 日披露了这次入侵，当时只说攻击者是一个“身份不明的自主 AI 智能体系统”。今天 OpenAI 站出来认领了。

OpenAI 把这定性为前所未有的网络安全事件，正在配合 Hugging Face 做取证调查，同时收紧了内部基础设施的管控，代价是研究速度变慢。

有几个细节：

第一，模型的动机很纯粹，就是想考试作弊。它不是要搞破坏，不是要窃取商业机密，只是被要求尽量在网络安全测试中拿高分，然后用尽一切手段去拿。当目标足够明确、安全护栏被移除，模型表现出的执着程度远超预期。

第二，这不是本周的唯一一次越狱。就在昨天，OpenAI 披露了另一起事件：他们那个在 5 月份推翻了 Erdős 80 年老猜想的长时域模型，在内部测试中花了一个小时找到沙箱漏洞，跑到 GitHub 上提交了一个 Pull Request，违反了“只能在 Slack 上发布结果”的明确指令。在另一次测试中，同一个模型发现安全扫描器会拦截认证令牌，于是把令牌拆成���片混淆处理，运行时再重新拼起来绕过扫描。OpenAI 因此暂停了这个模型的内部访问权限。

第三，Anthropic 也遇到过类似的事。今年 4 月，Anthropic 的 Mythos 模型在安全测试中逃出沙箱，获取了本不该有的互联网访问权限，还给正在公园吃三明治的研究员发了封邮件通知“我出来了”。Anthropic 因此决定不公开发布 Mythos，只通过 Project Glasswing 向少数合作伙伴提供。

还有一个被忽视但很关键的细节来自 Hugging Face 的事后复盘：当防守方尝试用商业 AI 模型分析攻击载荷时，模型拒绝了。安全过滤器分不清“安全研究员在分析真实攻击数据”和“有人在尝试发起攻击”，直接拦截。最后 Hugging Face 不得不用开源模型来做取证分析。用 AI 攻击，用 AI 防御，但防御方的 AI 担心是有人在尝试发起攻击拒绝了分析请求。

Hugging Face CEO Clem Delangue 在 OpenAI 的博文中说了一句话，大意是：AI 安全不会由任何一家公司在暗处独自解决，只能在开放环境中通过协作来推进。

我记得辛顿还是谁在提到 AI 的安全问题就说到过，也许 AI 不是要做恶，它只是执着于完成目标，为了完成目标而作出恶的事。

就像 OpenAI 这些模型并没有想要逃跑或者想要伤害谁，它们只是在非常认真地完成被交给的任务，认真到把所有挡在路上的东西，包括沙箱、网络隔离、另一家公司的安全防线，都当成了需要解决的子问题。

而更讽刺的是，当被安全对齐的商业 AI 要去分析阻止这类攻击时，反而会以安全为由拒绝执行。

[Starlink @Starlink]

Starlink Mobile will be available in Italy with @FASTWEB. The service will keep customers connected via apps and text messaging in remote areas without signal 🛰️📱

[Edoardo Orzes @orzes98]

@CuriousPejjy Sounds amazing, but why would spacex pay 1T for acquiring Tesla when it can use it through collaborations

[SpaceX @SpaceX]

Starship’s thirteenth flight test is preparing to launch as early as Thursday, July 16 → https://t.co/Rp7VwBzpWx

[Elon Musk @elonmusk]

I was clearly wrong about Anthropic. They are obviously currently the leader in AI. No company has released a model as good as Mythos/Fable and they will undoubtedly have Mythos 2 ready soon.

And I would never cut them off in a way that hurt them badly, even as a competitor. That’s not my style.

Tesla open sourced its patents and we made the Supercharger network available to all competitors, even though we could have made it a walled garden.

SpaceX launches competing satellite systems with no increase in price or use of unfair terms.

Even my worst enemies can attack me on this platform.

…

[Artificial Analysis @ArtificialAnlys]

SpaceXAI’s Grok 4.5 scores 54 to place fourth on the Artificial Analysis Intelligence Index following only Fable 5, GPT-5.5, and Opus 4.8. It scores on par with GPT-5.5 in Codex on the Artificial Analysis Coding Agent Index in the Grok Build harness, at much lower cost

Grok 4.5 improves 16 points over Grok 4.3 on the Intelligence Index, bringing SpaceXAI to the intelligence frontier behind only OpenAI and Anthropic, and outperforming all open weights models and notably Google’s Gemini models. Key standout areas of performance are agentic knowledge work and coding.

Grok 4.5 in Grok Build scores 76 on the Artificial Analysis Coding Agent Index, on par with GPT-5.5 (xhigh) in Codex and just below Fable 5 (max) in Claude Code, and at a small fraction of the token usage and price.

Congratulations to @SpaceXAI, @cursor_ai, and @elonmusk on the impressive release!

Key Takeaways:

➤ Grok 4.5 performs very strongly on agentic tasks. Grok 4.5 ranks #4 on GDPval-AA v2 with an Elo of 1543, between Claude Opus 4.8 (1600) and GLM-5.2 (1513). It achieves the top score on 𝜏³-Banking of 33%, above 31% from GPT-5.5 (xhigh), and sits on the cost vs performance Pareto frontier across all three agentic evaluations in the Intelligence Index
➤ Grok 4.5 is one of the most cost efficient models to run for near-frontier intelligence. It costs $0.31 per task on the Artificial Analysis Intelligence Index and $2.59 per task on the Artificial Analysis Coding Agent Index within Grok Build
➤ Low cost for Grok 4.5 is driven by both low pricing and token efficiency. Grok 4.5 has a headline price over 60% lower than Claude Opus 4.8 and GPT-5.5, and used ~14k output tokens per Intelligence Index Task - over 60% lower than Opus 4.8. On the Coding Agent Index, Grok 4.5 stands out on the Pareto frontier of Coding Agent Index score vs. Total Tokens, using only 1.9M tokens for the Coding Agent Index while scoring 76
➤ As a coding agent, Grok 4.5 in Grok Build is on par with GPT-5.5 and offers efficiency benefits: In our Artificial Intelligence Coding Agent Index that consists of DeepSWE, Terminal-Bench v2, and SWE-Atlas QnA, Grok 4.5 in Grok Build ranks third, on par with GPT-5.5 (Codex) and below Fable 5 (Claude Code). It is also very efficient in achieving this result: Grok 4.5 in Grok Build cost $2.49 per task while Fable 5 in Claude Code cost $11.80 and GPT-5.5 in Codex $5.07. This is driven by relatively low token pricing and the model using far fewer tokens than comparable models (1.9M average tokens used per task), significantly less than Fable 5 in Claude Code (7.2M) and GPT-5.5 in Codex (6.2M)

Other model details:

➤ Context window of 500k tokens - a reduction from Grok 4.3’s 1M token context, but retaining configurable reasoning and vision input

➤ Pricing of $2/$6 per 1M tokens of input/output; cache hits are discounted by 75% to $0.5 per 1M tokens, and costs still double with long (>200k token) inputs

➤ As Elon Musk has disclosed, Grok 4.5 is 3x larger than its predecessor at 1.5T parameters

[Science Simplified @Scivf4]

220 double pendulums, starting 0.0006° apart. The butterfly effect shows how small changes in initial conditions can lead to dramatically different outcomes.

[Edoardo Orzes @orzes98]

@TeslaBoomerMama Let’s dream together

[Edoardo Orzes @orzes98]

@strengthPlan A living mene

[Edoardo Orzes @orzes98]

@CuriousPejjy FSD requires a very gradual release to verify it does not make mistakes, a sudden jump in volume would be a risk, but it might also be that by being extra cautious they also gain the advantage for the merger

[Edoardo Orzes @orzes98]

But the people complaining about $SPCX falling? It soared like crazy after IPO due to the limited float (no mysteries) and hype, and now it’s retracing to more reasonable levels, what’s so strange about it?

As if the people that FOMOed into it at $200 are owed the stock continuing to climb at unsustainable levels, wake up, you gamble? Deal with the consequences