@vtxproject Those ready for intelligence-as-code must have a firm grasp on what they are trying to capture & how they wish to use it. This requires immense commitment to structured taxonomies, ontologies, classification for a knowledge domain. Without those, no tech will make a difference.

Many are quick to dismiss my ravings, knowing I'm a @vtxproject Synapse stan, but this isn't about a specific technology choice. If your thinking on the subject is reduced to graph tech and storage systems, then you are probably not ready for intelligence-as-code.

Unstructured and informal expertise is great, albeit ephemeral. Knowledge itself must be codified and hammered into your org at the institutional level, otherwise, it leaves your company with your human assets.

Lawyers and legal scholars tend to be some of the best writers in the world, probably because they spend so much time reading into the minutiae and dissecting the meaning of every little sentence.

RT @HostileSpectrum: The PDB (or any other finished intelligence) is not a palimpsest, to be rewritten to please. It is intended to be a ve….

"process engineering" is typically used in the manufacturing of physical products, but it can be applied for other domains. What do you manufacture? Do you know the flows of all the inputs and outputs, do you deeply understand the end-to-end process of making what you make?.

"be a factory, not a warehouse" -- if you build your knowledge graph with the right strategy and technology, it won't just store what you know, it will *manufacture* new insights. I believe this is true for the intelligence field, but this def applies to many other realms.

RT @notareverser: The inability or unwillingness of sandboxes to distinguish between network activity associated with the original file vs….

RT @cglyer: When I see discussion about AI in security (especially AI agents) - it’s often about how it will replace Tier 1 SOC analysts. I….

Those of you building modern edge devices, packet tools, network sensors, SSL decrypt, Suricata, etc -- it'll be a couple years yet, but your day will come again. Slow and steady like erosion, the attention, the investment, the market will come crawling back to you.

Imo the security product market is almost always a decade behind needs, but over time ends up being pulled to meet the adversary where they are operating. In the 2010s the market came late to the endpoint, in the 2020s late to the cloud, in the 2030s it'll be back to the network.

RT @hiddenlayersec: 🔐 Not all prompt injections are the same. We just released a taxonomy of adversarial prompt engineering, mapping the w….

RT @4a696d: So this one has been a while coming. If you've seen any of the talks or workshops I've done recently you'll have heard me talki….

vibe decoding.

RT @stvemillertime: @_devonkerr_ @HackingLZ Data breach got you down? Bring in your 1.05 material cybersecurity incident disclosure and get….

RT @anton_chuvakin: A fun #AI read:

My top 5 movies about ~hacking probably say more about my age than anything else, but still:. #1 - Hackers (1995) .#2 - War Games (1983) .#3 - Johnny Mnemonic (1995).#4 - Ghost in the Shell (1995).#5 - Office Space (1999) <- surprisingly full of hacks.

RT @stvemillertime: The Wire: Cybercrime leaves you feeling hard and cynical, aghast at a world embroiled in chaos, yet somehow also incapa….

Plot synopsis thread for the first five seasons.