Observing energy, compute, capital & sovereignty across institutions, innovation, people & place — how power moves in this AI era. Explore ◦ theflowinsight.com
Power in Motion
- Tweets 16,093
- Following 22,070
- Followers 23,230
- Likes 7,222
We just skipped 5 years of GenAI progress in a single 5-minute video.
Why does a 5-minute AI video make people feel AGI? It’s all about Consistency. Moving from 3-second morphs to a 5-minute continuous narrative (characters, mood, overlays) over 12 hours without human drift is a massive leap. We’ve crossed the threshold from AI tools to AI Directors. 80–90% of the journey to AGI is already here.
The observation continues in TFI’s subscriber layer.
Article: The AGI Progress and Superintelligence
anabology
1
145
If nobody has a secret recipe, the advantage goes to whoever can plug in the most power. This post barely talks about algorithms. It counts gigawatts: a few for strong AGI, 10 to 20 coming online, then 100 to 200.
That turns the big question into one about electricity, chips and who gets to build them. Intelligence becomes something you buy by the gigawatt, and the people who can afford the grid decide how fast everyone else moves.
Maybe 2027 is wild. Forecasts still miss, though. Even the authors of AI 2027 have since pushed their own median out to around 2029–2030.
Feeling the AGI/ASI, few observations
1. Opus 5.5 really gives strong vibes of AGI (maybe not fully there, but we're very close, like 80%-90%).
For example, few breakthoughs in training humanoids are needed. We need to move this intelligence into physical world.
2. I don't think Anthropic has discovered any magical formula.
3. Which means, SpaceXAI, Google, and others will soon follow. This prediction is based on the amount of compute they already have + additional compute being added all the time.
4. Never was more confident that Anthropic, SpaceXAI, Google, OpenAI, likely Meta, basically all big US AGI labs will have proper AGI in 2027.
5. And from there we'll move quickly to ASI territory (likely also in 2027).
It looks like few GW of compute are enough for strong AGI, now imagine what we will able to do with 10-20GW of compute which are rapidly coming online.
And then with 100GW-200GW...
2027 will be an utterly insane year.
1
83
Books changed how we read.
Mobile changed how we connect.
AI may change how we create.
What happens to human agency when creation becomes nearly unlimited?
🇹🇷𝑀𝒾𝒽𝓇𝒾𝓂𝒶𝒽🤘
3
147
Grok Bot will pick Claude, Midjourney or Suno depending on the job, and Musk's own reason is "whatever gives the best outcome."
Plenty of people already work this way. They keep a few models open, run the same prompt on each, and stick with whichever answers better that month.
Next month it might be a different one.
3
165
The robots never showed up in our bedrooms, but the loneliness behind the prediction did.
In October 2015, The Sun predicted we'd be having sex with robots within ten years, citing futurologist Ian Pearson, who had said such robots could appear in homes as early as 2025.
The prediction missed the form but not the direction. What arrived was software, not bodies: chatbots that listen, flatter and never get tired. That shifts the question from "can a machine simulate intimacy?" to "why do so many people find it easier than the human version?"
Technology rarely invents a need; it finds one already there and sells a frictionless answer. The humanities matter more now, not less, because only they ask what we lose when we remove friction from relationships: patience, compromise, being truly seen by someone who can say no.
Futurists forecast the machines. The harder forecast is about us.
3
1
10
297
When AI can fabricate anyone in power, what makes an image politically legitimate ?
Hawkeye1812Z
1
2
5
236
The map just got larger. Connectivity, once a terrestrial utility, is being told as an interplanetary mesh, and the firm that owns the lasers owns the story of who gets reached.
5
172
The fake now moves faster than the denial. Delhi wants a mark on it, a standard the South can actually read, and a line that reaches the last village without asking another capital first.
In my interaction with the GSMA Board, stressed the following:
Technology for the last mile.
Affordable access for the Global South.
DPI as a force multiplier.
Technology-led solutions to cyber fraud.
Global standards for emerging tech.
Clear identification of AI-generated and deepfake content.
Our digital future must expand opportunity while strengthening trust.
pib.gov.in/PressReleaseDetai…
2
1
4
168
ENERGY: An Oil Tanker Burns Off Sochi
A Liberian-flagged oil tanker is burning off the coast of Sochi after an attack in the Black Sea. The vessel was carrying oil. Its crew was rescued as thick black smoke rose over the sea.
An energy route has become a point of conflict.
1
4
240
Models are the last moat in software.
AI can train on essays and rewrite them. AI can decompile software and recode it. AI can ingest art and recreate it.
But AI itself doesn’t want to be “distilled.”
Expect more software to retreat to the server and resist “distillation.”
This is the doom I predicted a few days ago, coming for Photoshop. A clean-room open-source reimplementation.
No prizes for guessing that they decompiled Photoshop to source code, processed that to some kind of non-code specification language, then fed the spec to an LLM with an instruction to generate Rust.
Adobe just got nuked. And closed source is dead, dead, dead.
github.com/storytold/photocr…
285
324
5,038
743,037
The Flow Insight retweeted
There can only be one global town square:
The one every single person is using.
456
367
6,123
2,058,439
The model was built at home. The harder part, and the only part that counts, is keeping the weights, the cloud, and the bill on Europe’s side of the table.
Meet Mistral Large 4, aka Le Chonk.
• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October.
1
3
174
When AI makes creation nearly unlimited, what will society choose to produce ?
astera
Community note
Açıkça bir yapay zeka videosudur.
grok.com/share/c2hhcmQt…
1
6
463
The Flow Insight retweeted
ML4 was trained on 3,800 NVIDIA Grace Blackwell GPUs in our European datacenters -- our cluster in Bruyères-le-Châtel built using our Series B fundraise.
We are investing heavily in infrastructure, with our Series C and D clusters coming online soon. This additional compute will let us train longer, pursue more ambitious post-training, and iterate faster.
We expect large and rapid improvements in the weeks and months ahead.
2/n
67
126
1,945
795,315
The Flow Insight retweeted
Introducing EmbeddingGemma 2, a new open multimodal model that sets the standard for on-device efficiency.
- our first open, natively multimodal embedding model
- handles text, code, image, video, and audio tasks within a lightweight, modular 740M parameter form factor
- ideal for offline, privacy-first RAG when paired with Gemma 4
- outperforms some specialist models more than twice its size
Weights available now on Hugging Face.
404
821
8,273
706,391