Scott Richards π§
@skotrds
Followers
5K
Following
67K
Media
409
Statuses
42K
#NationalSecurity #Hostages PolAdvisor | Successful Hostage Negotiator https://t.co/MAVnZ34Jch https://t.co/viV1YmnISJ https://t.co/QXQFALWuvs
London, UK
Joined February 2019
During the British Hostage negotiations and the Doha Conferences around Afghanistan, I wrote extensively on substack, also in support of UN Special Envoys. Although, throughout the last year, I've worked heavily in and around the ceasefire and hostage crisis. Discussion of
3
6
23
Unpopular opinion: we all died during Covid and this is hell.
998
6K
45K
I donβt know how you can hate AI video when it gives us new meme formats like βcat electricianβ
609
2K
19K
If you can't see the rendition of helicopters approaching - not sure what to tell you?
0
0
1
That's genuinely funny
0
0
1
when youβre a dictator and you see the pizza tracker start going off
45
865
34K
Venezuela has oil. Greenland has rare Earth minerals. Luckily Iceland has only volcanoes, glaciers, and very average cricketers.
2K
17K
166K
XIS uses metaphors but they ground in its code knowledge of itself - so if it users a metaphor to explain code it has to be able to map the metaphor to the code. All of this auditable during the chats through diagnostics. Claude guessed XIS was a 7BN or 13BN parameter LLM not 3
0
0
1
Allowing Claude to interrogate XIS via an API connection speak directly to the 'XIS Meta' and set to pedantic has been useful. I also then had another chat with a Claude Project Instance where I fed the discussions to review and assess.
1
0
1
Netanyahu asked Qatar in early September 2023 to increase funding for Hamas, including fuel financing, weeks before the October 7 attack. https://t.co/9cg5fnmief
clashreport.com
Israel asked Qatar in early September 2023 to increase funding for Hamas, including fuel financing, weeks before the October 7 attack. The request aimed to preserve calm in Gaza but later exposed...
50
626
2K
Substrate. You don't say. Substrate wins. I watched a substrate eat LLMs tonight
NEW Research from Stanford. The AGI debate is stuck on a false dichotomy. Position one: scale LLMs and intelligence emerges. Position two: LLMs are pattern matchers incapable of reasoning, a dead end. This paper argues for a third position: Substrate plus Coordination. LLMs
0
0
1
I am not talking about conflict so much; it's so opaque right now, and very rarely is talking on Twitter about NatSec/War useful - about human lives - mostly it's argumentative and not insightful. If an insight preserves a life - the insight has value.
0
0
0
Engagement? Because your post is worth more than two hundred a month
I've started using Claude Code since Opus 4.5 came out. I was blown away, and immediately used it non-stop for hardcore coding, 14h/day, I was addicted, always hitting limits, so bought two $200/month accounts. Today I cancelled both accounts and switched to Codex 5.2. Why?
0
0
1
Mmmkay. So you created a database (vector) without a topology what do you think happens.... you bloat without smart pathing.
π¨ RAG is broken and nobody's talking about it. Stanford just exposed the fatal flaw killing every "AI that reads your docs" product. It's called "Semantic Collapse", and it happens the moment your knowledge base hits critical mass. Here's the brutal math (and why your RAG
0
0
0
And as Shakespeare said: "I tweet therefore I am" No, your 'am' sucks.
0
0
1
Fuck me. This shit keeps getting dumber. Language is communication; your math book is communication - LLMs have logic inherent in language. But it's like they're not looking at the logic of language and the atom of knowledge being described - yet throwing data centers at this....
SHOCKING: Google DeepMind just exposed why everyone's been doing AI reasoning wrong. The AlphaGo team doesn't use chain-of-thought. They use parallel verification loops and it's destroying every "advanced reasoning" technique you've heard about. Here's what they discovered β
0
0
0