Pinned Tweet
The Founder-Parent Paradox:
You're building a business to provide a better future for your family, but the act of building it takes you away from your family in the present.
The ultimate balancing act.
I tested Recursive Flint (40B, 2.5B active) on one DGX Spark: 93 tok/s, 2x the speed of Qwen3.8 Flash Next, in a third of the memory.
Best IFEval I've measured (94.6).
But it quits early on long tool chains (multi-turn BFCL 32 vs 54.5) and thinks Kortrijk is the capital of West Flanders. 🫠
For now, Qwen stays.
ALT Bar chart, one DGX Spark: Recursive Flint 40B-A2.5B vs Qwen3.8 Flash Next FP4-Local. Tool calls 80.0 vs 82.4, multi-turn tools 32.0 vs 54.5, agent tasks done 75.0 vs 84.4, attacks resisted 94.5 vs 99.0, IFEval 94.6 vs 85.8, Dutch knowledge 79.3 vs 87.0. Decode 93 vs 44 tok/s, agent loop 24.8 vs 13.8 steps/min, memory 33 vs 97 GB.
Docker Compose setup, tuning and the full results: github.com/dedene/Recursive-…
Look. I don't want to overstate it. But when you look at contemporary AI Twitter — the cults of AGI timelines, consulting an oracle that answers in confident riddles, endless debates in the agora about benchmarks — it's hard not to notice some parallels with ancient Athens.
when anthropic's CISO says spend it on the basics and not on ai, maybe listen 👀
In hindsight this makes total sense, but I didn't see Canva training their own imagegen model coming 🔥
Introducing Canva Craft 🚀
Extraordinary image generation is now available to everyone, powered by our first frontier model built specifically for design.
With Craft, Canva is now the world's second most advanced AI lab for image generation on @DesignArena's global leaderboard.
See what it can do 👇
Asked Grok to help figure out why I got blocked when all our interactions were friendly.
Even Grok came back with "I truly have no idea" 🙃
Jensen Huang telling little robots to "take a nap" on stage at GTC Berlin is peak 2026.
They're from Pollen Robotics, now part of @huggingface, and he says they learned their skills in a digital twin on an RTX Spark.
Who of you ordered one? 👀
Think about it. An LLM trained in 1789 under the Ancien Régime would have told you democracy was "a dangerous idea". Anno 2026 nobody questions it anymore.
That's a danger I think about a lot. Someone sets the baseline of what counts as "acceptable thinking", and right now that's only a handful of labs.
Not every guardrail is censorship. Some are legal cover, or protecting us against real dangers, and that's fine. But one moment's values get baked into a tool many of us work with all day long.
Open weights didn't solve this either. Releasing a model and having a real alternative are two different things.
That's what I like about the Distributed Cognition bet: superintelligence doesn't need to be one giant model. It can be a network of specialized ones all talking to each other.
Nature did it that way too. Young chimpanzees can beat us on some short-term memory tests. But we got language instead, and that lets our limited minds wire together and create beautiful things.
Funny thing, good agent setups already looked like this. Not one prompt doing everything, but a team of agents with narrow jobs, passing work back and forth.
If you don't own your intelligence, you're renting your thinking.
To me, it's worth knowing who your landlord is 👀
🤝 Paid partnership
Today distributed cognition exits out of stealth with backing by Verdict Capital.
Our mission is to replicate humanity's collective intelligence.
Read more: dcog.com/research/our-impera…
Most of today's AI power is held in the hands of a few, who can choose what you're allowed to think and what you're permitted to do.
We stand on three substantially different foundational beliefs:
- You shouldn't ask permission for intelligence.
- Superintelligence will be a distributed collective.
- Humans are not meant to be relegated to a permanent underclass.
If we don't distribute intelligence, nobody will.
Sam Altman told @MatthewBerman he has "not read or even looked at a line of code in probably a year."
If he ever writes code by hand again, it'll be "like a revolutionary war reenactment" 💀
Funny line, but somebody still has to read what the agents ship.
He also hates the usual "AI is going to cure cancer" pitch. Dig a little deeper and the price is often "human agency and you in control of the future."
What he hopes for instead is that we "avoid this industrial revolution trap" and give people more choice, freedom and time.
RSI, consciousness, the cancer pitch.. his biggest answers all came back to who stays in control. That's a fair test for the agents we ship too.
Full interview by @MatthewBerman with @sama: youtube.com/watch?v=3gXLmy7D…
This week is BRUTAL for devs everywhere.
Monday vs today 😭
Working RCE against PHP 8.5.12 (the 2026-10-21 security release that fixed CVE-2026-109618/109619).
Pure logic bug — no memory corruption,
ASAN-clean, orthogonal to the CVE-2026-109618/19 class.
Receipt: uid=33(www-data) over HTTP, one-shot
#infosec #php2shell #php
Readers added context they thought people might want to know
The claimed RCE on fully-patched PHP 8.5.12 requires a custom lab app with a destructor gadget plus deliberate session.serialize_handler mismatch (php_serialize writes, legacy php reads) and non-default settings like cleanup=Off; this is an app flaw, not core PHP RCE.
github.com/rafalohaki/php…
php.net/manual/en/sess…
It just.. keeps.. coming 🤯
is it time to turn the internet off yet
kb.isc.org/docs/cve-2026-943…
