🌚 YRA 🌝
6.7K posts
טורקיה תתקוף את ישראל? ניתוח גיאופוליטי קצר -
הסבירות: פחות מ-2%. הנה למה:
🔹 נאט"ו — התקפה על ישראל = משבר קיומי לברית. ארדואן לא אובדני.
🔹 כלכלה — אינפלציה, הלירה מתרסקת, הוא צריך השקעות זרות — לא סנקציות.
🔹 צבא — ישראל עם יתרון איכותי עצום, הגנת אוויר מהמתקדמות בעולם, ועמימות גרעינית.
🔹 ארה"ב — תתערב. זה לא "אולי". זה כמעט ודאי.
אז מה ארדואן כן עושה? רטוריקה, תיאטרון דיפלומטי, תמיכה בחמאס, הגבלות מסחר. מעצבן? כן. מלחמה? לא.
התרחיש היחיד שמדאיג: הסלמה ספירלית בסוריה שגוררת שתי המדינות בטעות, או משבר פנימי שדוחף מנהיג נואש ליצור אויב חיצוני.
שורה תחתונה: תעקבו אחרי מה שהוא עושה, לא מה שהוא אומר.
עברית
נשיא טורקיה, ארדואן אומר שאם המלחמה תתארך והמשא ומתן בין ארה"ב לאיראן בפקיסטן לא יוביל לשלום, ייתכן שטורקיה תיכנס למלחמה נגד ישראל.
עברית
What does the MMIWG2SLGBTQQIA+ flag look like
English
That’s true, bad code is still bad code no matter where it runs.
But I think that’s narrowing the argument a bit and missing the bigger point.
A lot of real world exploits don’t come from the code alone, they come from everything around it:
APIs, misconfigurations, exposed endpoints, cloud networking, permissions, etc.
Traditional stacks are full of connection points that all introduce attack surface.
What Dom is getting at is reducing that surface at the architecture level.
If you run everything in a unified, tamper-resistant environment, you eliminate a lot of those external risks and can focus more on writing secure code instead of securing the infrastructure around it.
It doesn’t magically fix bugs, but it removes a huge amount of complexity and potential failure points. Making it easier to secure.
$ICP
English
We are reaching the point where software MUST be *tamperproof* — hosted by secure cloud infra created by BFT networks, where security derives from math, not from patching holes, because AI can find the subtlest of flaws. This is what Internet Computer technology does TODAY.
Nina Schick@NinaDSchick
Claude Mythos. Ten trillion parameters: the first model in this weight class. Estimated training cost: ten billion dollars. On the hardest coding test in the industry (SWE bench) it scores 94%. It found a security flaw in a system that had been running for 27 years, one that every human engineer and every automated check had missed. It found another bug that had survived five million test runs over 16 years. (It did so overnight.) It is so capable in cybersecurity that Anthropic will not release it to the public, instead it is launching Project Glasswing along with 100m in compute credits to help secure software. Only twelve partners currently have access: Amazon, Cisco, Apple, Google, Microsoft, NVIDIA, JPMorgan Chase, Crowdstrike, Palo Alto, AWS, The Linux Foundation, Broadcom. (I'm sure the Pentagon is on the line?) This is not a product launch: it is a controlled deployment of a system too powerful to distribute freely. Tell me this isn't (very expensive) AGI?
English
You're right that ICP removes traditional infra vectors. Nobody's arguing that.
But Ethereum and Solana also run code in secure virtual execution environments on BFT networks. Billions in exploits anyway. The Wormhole hack, the DAO, Mango Markets — all application logic bugs, not infrastructure breaches.
A mathematically secure network doesn't fix a logic error in your canister. That's what Mythos finds.
English
@ICPsimp @YRA_IL Not possible on Internet Computer in traditional sense: APIs, misconfigurations, exposed endpoints, cloud networking, permissions, etc.
Software runs inside a scalable virtual execution environment inside a mathematically secure network, not on "infrastructure" can escape into
English
Fair points honestly. The logic of reducing attack surface to force attackers into fewer paths is legit.
Just don’t oversell it. There are still physical nodes in data centers, just managed differently. Immutable code is a double-edged sword when you ship a bug. And phishing and social engineering cause most breaches — ICP doesn’t touch that.
Infrastructure-level improvements matter. But most real-world damage happens at the human layer, not the infra layer.
English
$ICP doesn’t solve every single issue.
But that’s not really the point.
The point is reducing the attack surface and removing entire categories of problems, not pretending anything is 100% secure.
Right now most systems are wide open at the infrastructure level.
Things ICP largely eliminates:
Server/infrastructure attacks- no servers or OS access
Misconfigurations - no AWS-style setup to mess up
Single point of failure - no “hack one machine/account - control everything”
Silent code tampering - deployed code is immutable
Things it greatly reduces:
Key theft - threshold signing, no single key exposed
Supply chain risk - immutable code + multi-approval updates (Orbit)
Data breaches - no centralized database to dump
Account compromise impact - can require multiple signers
Still possible:
Bad code
Phishing / social engineering
Poor governance decisions
Right now cybersecurity is basically being fought from every direction, and resources are stretched trying to cover everything, so things slip through.
If you reduce the attack surface like this, you force attackers down fewer paths, which makes it a lot easier to focus defenses and actually secure what’s left.
English
"The next wave of AI-powered cybersecurity attacks will be like nothing we’ve seen before."
Internet Computer technology creates tamperproof clouds using mathematically secure networks.
cnn.com/2026/04/03/tec…
English
איזו תמונה מטורפת עם טוויסט מגניב:
היא צולמה בלילה.
כלומר, השמש נמצאת מאחורי כדור הארץ.
אז איך בכל זאת רואים משהו?
כדור הארץ מואר על ידי הירח - והתמונה צולמה בחשיפה ארוכה.
מישהו הצליח להבין איזה חלק של כדור הארץ רואים?
עברית
@paoloardoino Deloitte, EY, KPMG, PwC… all gonna get rekt eventually.
English
תזכורת שזה היה יכול להיות סוורוס סנייפ, אבל HBO היו חייבים להרוס את אחת הדמויות הכי חשובות בסדרה.
עברית
64.87% of the $ICP Staked supply is locked for 8 Years.
With M70, Neurons that have historically maintained an 8-year dissolve delay will be given a special flag and receive a dedicated reward boost of 10% in recognition of their long-term alignment with the network.
The reward boost would remain in effect until the end of 2030. This special flag, and the corresponding reward boost, would be lost once the neuron starts to dissolve.
M70➡️ new maximum dissolve delay = 2 years
English
@CryptoWendyO USDC is crushing because of news that Tether is being audited by the Big Four.
English
Defi is done.
CBDC incoming.
Watcher.Guru@WatcherGuru
JUST IN: USDC stablecoin issuer Circle $CRCL stock crashes 18% today.
English
תזכורת: המדע אפילו לא יודע להסביר למה מגנטים נמשכים זה לזה
עברית
🚨M70 UPDATE:
Yes, some of you noticed this, the voting time was extended from 15 hours left to 1 day and 11 hours because of the “wait-for-quiet” mechanism.
The idea is simple: if everyone agrees, proposals pass quickly. If things are controversial, more time gets added so people can keep voting.
Every time the result flips (from “Yes” to “No” or the other way around), the deadline is pushed further out. The more contested it is, the longer it stays open.
The base voting period is 4 days, and it can be extended by up to another 4 days.
$ICP = World Computer ♾
English