Living the mediterranean life
Malta
Joined January 2012
- Tweets5K
- Following254
- Followers487
- Likes12.9K
map/reduce retweeted
Where does your AI agent's reasoning happen?
For a lot of agent tools the answer is someone else's cloud. Lascaris runs in your cloud account, on-prem or air-gapped. Reasoning never leaves your network.
scalytics.io/lascaris
Rules and alarms run inside infinimesh core.
A rule is a CEL condition on device state, like reported.temp > 70. When it fires: a log line, a signed webhook, an MQTT publish, or a patch to the device's desired state. Alarms go raised, acked, cleared, with cooldown and repeat.
github.com/infinimesh/core/b…
Patch released last weekend. Core, the inf CLI and the operator are on v3.4.4, addons on v3.4.3, all built with Go 1.27.3 for CVE-2026-111943 in crypto/x509. If you run 3.4, upgrade.
Want a first PR? Issue #77 asks for a curl walkthrough of rules and alarms against a local stack.
github.com/infinimesh/core/i…
Anthropic, Google and OpenAI now mark AI output. Some of it is metadata. The hard part is statistical: the mark sits in which words the model picked.
clean-ai-marks handles both:
- rewrite with a different model (vLLM, Ollama, any OpenAI-compatible endpoint) to scramble the word statistics
- strip C2PA, EXIF, XMP and doc properties from PNG, JPEG, SVG, PDF, DOCX, ODT, HTML, Markdown
- remove zero-width, bidi and tag characters
- inspect, clean, inspect again
Stdlib Python, runs as an agent skill. Statistical marks get reduced, not provably gone.
github.com/novatechflow/clea…
map/reduce retweeted
‼️ Researcher Erica Windisch skipped coordinated vulnerability disclosure and published 11 unpatched Tailscale flaws.
In her lab, Tailscale SSH race conditions let one hacked device get root on other servers and add a rogue node.
Readers added context they thought people might want to know
The Tailscale SSH findings require prior authorized access on the tailnet per the researcher's report; Tailscale stated issues concern post-authorization behavior only and did not recommend disabling SSH for everyone.
tailworm.com
x.com/IntCyberDigest…
map/reduce retweeted
Im speaking at Community Over Code Sydney 2026, 18 to 19 November on Groovy 6 features for humans and AI agents, but there are 30+ other talks on technologies that will impact our futures. Early bird tickets finishing soon!
communityovercode.apache.org…
@TheASF @Chur_Conf
map/reduce retweeted
More private keys by the Indian certificate authority have been discovered. These are all small numbers.
People are suspecting that they interpreted 384 bit keys to mean that they should pick numbers from 1-384 to use as their private keys.
map/reduce retweeted
The fun fact about the eMudhra certificate leak is that the numbers were so low that only an AI or a flawed/test script would have found them. No sane human would have thought that a fellow human could have done something that stupid.
If you run php, you’re hacked. If you installed vllm latest even, hacked. Node.js? Hacked. Indian SSL certs? Yeah you know. And it’s just Wednesday.
map/reduce retweeted
Can you trace what your agents saw when they got it wrong?
scalytics.io/check
Google Cloud down in Europe, and with it quite the largest one from finance, industry and manufacturing. Now imagine when the US puts a stop to all US-based hyperscalers? It's time to work together, again!
map/reduce retweeted
Your Kafka is replicated. Is it backed up?
Delete a topic and your standby deletes it too. Two clusters, no data.
4 tests a real backup must pass: scalytics.io/kafka-backup-vs…
map/reduce retweeted
Every AI agent action should answer 4 questions:
What did the source say?
What is true enough to decide now?
May this decision be made here?
What changed, and who allowed it?
IBM + Confluent Kafka is where those answers live: scalytics.io/blog/ibm-conflu…
Uhhh - now it gets interesting. Cryptographic curve seed discovered.
CLAUDE FOUND THE SEEDS
anthropic.com/research/ecdsa…
map/reduce retweeted
NEW VIDEO🔥 Dear ppl who hate America…
DOWNLOAD SONG: itunes.apple.com/us/album/wh…