
Tweeting about #Finance #Banking #WealthManagement #Greece #Luxembourg #Fintech
Luxembourg
Joined May 2010
- Tweets886
- Following305
- Followers53
- Likes89
Interesting but very looong read (40000 words) about #AI. Better to listen to it...
commanding-heights.com/
In case you don't have time for that you could watch the AI generated video below
teapawt.net/cdcwatson/status/21172…
1
4
Damodaran on #Anthropic IPO valuation: The Anthropic Prospectus Arrives: My (but not the) IPO valuation! aswathdamodaran.blogspot.com…
10
Very interesting, opinionated article about deal clustering (#Anthropic and #Facebook)
giftarticle.ft.com/giftartic…
6
Food for thought on #AI: historical parallels and the possibilities it could unlock contraptions.venkateshrao.co…
9
ICYMI: OpenCode is on a roll with free powerful models being offered on a daily basis.
Yesterday: DeepSeek V4.1 Flash
Today: GLM 5.3 Flash
opencode.ai/v2/docs/console/…
24
A Change in AI Strategy: turning points -Luna repricing, Haiku 5.5 and GrokBot decision to be a distribution channel. Apart from the ongoing pricing war, distribution castles are rising. The Ultimate moat? tomtunguz.com/what-if-the-mo…
8
Intersting report by epoch.ai: How do Chinese AI companies make money? epoch.ai/publications/how-do…
3
New A16z State of Markets II report doesn't disappoint. From all the interesting graphs though, I chose the last: Collect them all :)
1
The rise and importance of system one or decision models (Jev and lookalikes)
How System One Models Like Jev Change Enterprise AI Architecture kai-waehner.de/blog/2026/09/…
7
"The ugly economics of consumer AI put a hard cap on how large the company can plausibly grow without tapping into enterprise revenue. It’s a lesson the major labs have already learned, and it’s one of the thingsthat doesn’t seem to be changing"
techcrunch.com/2026/09/30/th…
2
Towards #RSI (Recursive Self Improvement) at #Anthropic
26% of AI R&D currently is "lead" by AI itself -> Interesting
More interesting: That figure was 1%, 6 months ago!!! anthropic.com/institute/meas…
6
Is Hong Kong market simply weak, or has the torrent of listings overwhelmed the available liquidity?
Could the same dynamic be unfolding globally?
The demand for AI investment is absorbing vast amounts of capital, sucking up the marginal capital
giftarticle.ft.com/giftartic…
4
Another day, another AI Capex forecast. This time from UBS
YES AI capex is rising but if you look closely the projected numbers, the interesting points are different:
No access to the original report but IMO the important points are badly reported
finance.yahoo.com/technology…
2
OpenClaw for Normies? Meet Instinct.No technical setup.
Get a glimpse of what an agent is and how it can ACT.
Still in beta, still imperfect but still free.
DM me if you'd like an invite
aibyaakash.com/p/instinct
app.instinct.co/
1
114
A major transaction was signed by Sequence holdings. But how Sequence characterizes itself: Refounding: rebuilding an established enterprise around what is now possible with technology – its operations, its workflows and the potential of its people seqholdings.com/blog/sequenc…
6
Astra’s 99.9% score on ARC-AGI-3 was achieved using a custom harness. Score without it: 62.7%. [arcprize.org]
So, even without training models, there is room for improvements: harnesses, context, efficiency not to mention fine-tuning.
So, what are we discussing about?
8
TL;DR: Allocate €100b to AI infrastructure, (data centres and compute). Help attract €1.5tr in private investment -> Increase Europe’s share of global AI compute from 5% to 15% by 2030.
Failing to act, risks creating permanent dependences.
transformative-ai.eu/part-a#…
6
The same week that OpenAI admits it “cannot rule out that de-identified data derived from usage of our products helped improve our models”, it releases Defense Factory
Your best practices – our training data
openai.com/the-defense-facto…
4
Matemathicials used ChaTGPT.
OpenAI might have benefited from those "de-identified" chats and used them to solve the problem.
“While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models." techcrunch.com/2026/09/08/op…
17


















