Pretty great
All can be true
Interesting to think as models & harnesses keep improving, which pieces become implementation details that are the agent's job, and which abstractions stay
Bar goes up, not down
Before joining a team or starting a project, it's important to know which tribe of agentic engineering you're choosing:
- The software engineers: Treat agents like another software engineering problem. Borrow ideas like Map-reduce to enable parallel agent swarms. Build
GFQL -- open source CPU/GPU cypher graph queries on dataframes, no DB needed -- keeps getting better!
Indexes aren't just for databases. This GFQL release further increases our CPU+GPU mode speedups over neo4j, memgraph, & kuzu on even more benchmarks, both for low-latency
Your dataframes can now anwser graph queries in milliseconds, without a graph database.
A Cypher query across 50 account IDs went from 2.7 seconds to 5 milliseconds. Searching 30M timestamps went from 40 seconds to 101 ms.
The latest GFQL release adds property indexes, so you
DM if we're intersecting to chat agentic investigations or graphs!
Also not listed, I'll be in San Diego and then DC/VA/MD for a couple days each
Exciting to be able to share more of what we've been up to :)
We’re about to spend most of October on the road, and we’d like to meet you at one of these stops.
Here’s where we’ll be:
Oct 17: BSidesNYC, New York
Oct 19 to 24: BSidesAugusta week, Augusta, GA
Oct 27 to 29: [un]prompted II, San Francisco
Oct 29 to 30: LASCON, Austin
Augusta
Latest botsbench on Astra vs Fable 5.1 is out! Astra is quite the leap, though still better to do an investigation harness on older models than using a coding model and (incorrectly) hoping that paying more wins
Also, newest D&D model cards: both got a Lawful Neutral rating,
GPT-6 Astra and Claude Fable 5.1 are out, and they're pretty compelling if you have the budget.
But put them into a real investigation harness, and things get interesting.
On CyBT-CTF, 59 incident-response cases worked in Splunk, Astra reaches 69% and Fable 5.1 reaches 61%.
