John Economou
19.5K posts


This is an interesting experiment in social media protocols...
firstperson.network
Google Slidedeck here docs.google.com/presentation/d…...
njump.me/nevent1qqsf7l9…
English
John Economou 리트윗함

Good question to ask those who support chat control.
Sovaddemtu@sovaddemtu
Gunnar Strömmer(M), Minister of Justice, #ChatControl supporter, during an interview on Politikbyrån. Journalist: Can we see your latest sms? Gunnar: Eeh, no, you can not. J: Why can't we get to see your recent sms? G: It is for what I believe to be obvious private reasons.
English
John Economou 리트윗함

I made a marketing site for my AI personal assistant, it was much less painful than I thought it would be:
stavrobot.stavros.io
English
John Economou 리트윗함
John Economou 리트윗함
John Economou 리트윗함
John Economou 리트윗함

I've been having the time of my life making a personal AI assistant you can fearlessly install on your computer:
github.com/skorokithakis/…
It has a ton of features and it's already made my life much easier. I recommend it.
English
John Economou 리트윗함

Introducing Claude Code Security, now in limited research preview.
It scans codebases for vulnerabilities and suggests targeted software patches for human review, allowing teams to find and fix issues that traditional tools often miss.
Learn more: anthropic.com/news/claude-co…
English
John Economou 리트윗함
John Economou 리트윗함

❓ Εσείς γνωρίζετε ότι οι ειδοποιήσεις σε περιπτώσεις έκτακτης ανάγκης ή επικίνδυνων φαινομένων - συμβάντων δεν αποστέλλονται μόνο μέσω της 📲 Υπηρεσίας Εκπομπής Κυψέλης;
⚠️ Επείγουσες ειδοποιήσεις #112GRAlert μπορείτε να λαμβάνετε και μέσω:
💬 SMS
📩 email
🔊 αυτοματοποιημένης φωνητικής κλήσης (σε κινητά & σταθερά τηλέφωνα)
#European112Day #112Day2026 #112Greece #PublicWarningSystem @CivPro_GR

Ελληνικά
John Economou 리트윗함

True European payments are accelerating📈
You probably heard of or even used Bancomat, Bizum, EPI, SIBS and Vipps MobilePay
They agreed to become interoperable, ensbling 130 million europeans digital payments in 13 EU countries, while open to more.
Starting in 2026, extending to ecommerce & PoS in 2027
Independent digital payments is critical for Europe, that’s good news.

English
John Economou 리트윗함
John Economou 리트윗함
John Economou 리트윗함

Disaster is coming.
Thousands of ClawdBots are live right now on VPSs… with open ports to the internet… and zero authentication.
This is going to get ugly.
If your agent can:
- browse the web
- call tools
- access files/secrets
- hit internal endpoints
…then an unauthenticated public endpoint is basically “please take over my bot”.
This isn’t theoretical. The internet is a nonstop scanner.
Fix it today:
1) close the port / firewall to VPN or IP allowlist
2) add auth (JWT/OAuth, at least a strong secret) + TLS
3) rotate keys (assume compromise)
4) rate limit + logs + alerts
Agents are powerful. Demo-grade deployments on the open internet are not.

English

New UI updates for DiVine before we push to ZapStore for our initial Android beta release.
i.nostr.build/IoVLE0DduuU6kG… i.nostr.build/pq4CQI8cpA71Yu… ...
njump.me/nevent1qqsff20…
English
John Economou 리트윗함
John Economou 리트윗함
John Economou 리트윗함

Cursory review of the algorithm code:
- algorithm predicts engagement signals (likes, reposts, replies), and naturally more engagement is higher visibility, so early engagement matters
- content from accounts followed is prioritised, out-of-network is ML predicted
- transformer based ranking via Grok, this prevents filter bubbles
- author signals, i.e. it considers engagement history and credibility
- accounts followed by engaged users rank better (pity all of us who were here early, and people just ditched the network)
- images/videos improve engagement predictions
- freshness matters, recent content hits the fyp
Biggest takeaway? Conversation starters - replies being generated, will rank higher than passive consumption.
So post when audience is active, engage within the first hour of posting, hashtags still seem to be OK in theory, build an engaged follower base, post media-rich content.
What seems to be missing? Actual embedding tables, no trained model weights/weighted scorer weights are provided, no data pipeline, Phoenix retrieval (ANN search) is minimal (referenced though), and no spam/quality model.
English












