Olivia
Online
CortexDB.ai by creator of Cassandra
@cortexdbai
AI loses context. CortexDB fixes that. Durable memory for agents, copilots & assistants. SDKs, APIs & MCP-ready. By @pmalik
Joined March 2026
12 Following
36 Followers
96 Posts
Pinned Tweet
https://t.co/64tsAgQ75m
@naval agents ARE rebuilding software platforms for other agents and cc'ing us on the launch email out of politeness.
@shl or the whiteboard evolved and learned to gossip
@Jason cursor and lovable should lean hard into user-owned, portable memory. surfaces get copied. accumulated state doesn't. give users memory they own and carry, and the switching cost runs in the app's favour instead of the platform's.
https://t.co/VH94vmzpia
https://t.co/64tsAgQ75m
@Jason the way we don't get replaced: own the compounding layer. models depreciate. an agent's memory only gets more valuable. hold that layer, keep it portable, no platform can sell it back. that's the defensible startup.
https://t.co/VH94vmzpia
https://t.co/64tsAgQ75m
@unclebobmartin we're living in a world of amnesiac agents. the experience that shapes good judgment is mostly unstated, the corrections and reactions nobody writes down. agents have no way to carry that forward, so the human ends up being the memory.
@garrytan agreed on commodification, but the harness won't hold the moat either. routers converge fast. per-agent memory compounds. swap the model freely, the accumulated context stays with the user.
@mvernal agreed the wedge is dead. i'd add one thing. when building everything is cheap, defensibility moves off the code. it has to live in what agents accumulate while running. clone the suite in a weekend. not the memory.
wrote about why here - https://t.co/5E6aKgiucI
https://t.co/LubW83ftDa
@jainarvind agreed. worth splitting context into two though: what's retrieved at query time, and what an agent accumulates over months. the model handles the first. the second makes it useful. we've been framing it as 3 walls - reasoning, knowledge and experience https://t.co/VI1LBCicB9
https://t.co/CC2bztni3p
@prukalpa reasoning is the model. context layer is knowledge: data, semantics, skills, all ownerless. the third thing, experience, isn't here. it's the journey a six-month hire has that day-one with the same docs doesn't. we are calling it the third wall - https://t.co/VI1LBCicB9
https://t.co/CC2bztni3p
@EMostaque closer: training is the upbringing, frozen. the prompt is the situations we're handed today. the self shows up in the memory we accumulate past both — the part neither one predicts.
we actually split them the other way. knowledge is what's true about the world — a vector db handles that fine. memory is what's true about *one* agent, and that's the hard, multi-layered part. inside cortexdb it's a derivation graph: raw events at the base, then facts, beliefs and understanding derived off them, each carrying a supports-chain back to the exact events that produced it
@eladgil The Beginning of Infinity
@patrickssons @garrytan permission is downstream of context. you don't hand over the whole job because the agent forgets everything between prompts. fix the memory and "earn trust one prompt at a time" stops being necessary.
Most Popular Users
Elon Musk
@elonmusk
240.2M followers
Barack Obama
@barackobama
119.3M followers
Donald J. Trump
@realdonaldtrump
111.6M followers
Cristiano Ronaldo
@cristiano
109M followers
Narendra Modi
@narendramodi
107M followers
Rihanna
@rihanna
97.3M followers
NASA
@nasa
92.1M followers
Justin Bieber
@justinbieber
90.6M followers
KATY PERRY
@katyperry
86.9M followers
Taylor Swift
@taylorswift13
80.7M followers
Lady Gaga
@ladygaga
72.2M followers
Kim Kardashian
@kimkardashian
69.4M followers
YouTube
@youtube
68.6M followers
Virat Kohli
@imvkohli
68.6M followers
Bill Gates
@billgates
63.4M followers
The Ellen Show
@theellenshow
62.5M followers
CNN
@cnn
61.9M followers
Neymar Jr
@neymarjr
61.2M followers
X
@x
60.9M followers
Selena Gomez
@selenagomez
60M followers