Geopolitical and macro cycle analyst. Founder @NoodlesResearch. I connect the dots others miss. -The Research Terminal-
- Tweets 23,323
- Following 2,580
- Followers 106,391
- Likes 6,118
Pinned Tweet
On January 9th, we wrote that Greenland was theater. That the real target was Iran. And that oil would cross $100.
Brent was trading below $70. Nobody was talking about Tehran.
Five months later: the largest air campaign in Middle East history, the Strait of Hormuz closed since February 28th, Brent averaging $107 in May.
And today, Trump announced the next step on Truth Social: “We will be taking Kharg Island, and other oil infrastructure points, and assume total control of their Oil and Gas Markets.”
Now read what we published in The Iran Playbook, days before his announcement:
“Phase 3: The ‘Limited’ Operation. Seizure of Qeshm Island. Possibly Kharg Island. Framed as a maritime security operation. ‘Protecting critical energy infrastructure.’ Never called an invasion.”
Word for word. The island. The framing. The vocabulary.
This is not luck. It is pattern recognition. Every American war follows the same script: a provocation, a “defensive” response, a “limited” operation that becomes permanent. Vietnam started with advisors. Iraq started with liberation. Iran is starting with an island.
An occupation that cannot be defended without escalating, and cannot be abandoned without humiliation.
This is the moment the world change. The socio-economic shockwaves of a permanent war in the Gulf will hit every economy on the planet: energy, food, shipping, inflation, debt.
America is dragging itself into a vortex with no exit. And it is dragging everyone else down with it.
15
7
77
40,489
Half the people making $100k+ are terrified of losing their corporate jobs.
The government calls stepping away "Not In Labor Force" (NILF).
We just swapped pointless Zoom meetings for 3 screens in the trenches.
That’s why I think robinhood:0xdb97b2aefd6816485a757ae49280d598a5c9c848 has the most potential on Robinhood Chain right now. It’s not some random forced mascot, it’s literally our daily reality.
$NILF
Americans are increasingly worried about their jobs.
The perceived probability among US consumers that the unemployment rate will be higher in 1-year increased +1.6 percentage points in August, to 44.4%, its highest since April 2020.
Excluding the 2020 pandemic, this marks the highest reading since the New York Fed Survey of Consumer Expectations began in 2013.
By comparison, this indicator peaked at 44.1% in April 2025, following uncertainty around tariffs.
By income, the perceived probability of a rise in unemployment was the highest among those earning over $100,000, at 50.3%, the 2nd-highest reading on record.
By comparison, expectations stood at 43.5% among those earning $50,000-$100,000, the 2nd-highest reading since March 2025, and 38.0% among those earning under $50,000, the highest since May.
Fear of unemployment is rising.
29
14
58
10,423
No, Trump and his elite will never leave the throne; one way or another, Trump will get a third term as well.
Trump and JD going to get smoked in the midterms.
The Clarity act vote was a precursor to the results.
Trump has aligned himself with Crypto and they will use Crypto as the cudgle to go after him.
Trump will spend the next 2 years nuking the USD, YCC to get bonds down because there is nothing else they can do.
Risk assets go up
6
17
8,896
This isn't an administration; it's a farce. And it’s not even funny, because it is about to shatter the entire global economy and drag the US down with it.
5
46
7,819
So ‘Jerome Warsh’ hike interest rates…
So Jerome Powell was just the scapegoat.
On January 9th, we mapped the chain. Iran → oil → inflation.
Today the Fed’s preferred inflation gauge came in at 3.8%. Up from 2.4% in January. At this pace, June prints above 4%.
Now the second chain is playing out.
Persistent inflation → Fed trapped → liquidity compression → markets under pressure.
Trump picked Warsh because Powell wouldn’t cut fast enough. Now the market is pricing a 50% probability of a rate HIKE by year end.
Trump started a war that sent oil from $65 to $100+. That inflation made cuts impossible. His own Fed chair might have to raise them.
He built the trap. Then walked into it…
16
7,627
Still relevant.
Its just funny everyone spent months blaming Pumpfun for saturating the market and killing the trenches.
Now, in full bear mode, your big new narrative is recycle launchpads. And every pitch is “the next Pumpfun” on every chain.
The thing that killed the trenches is now your bull case.
3
17
8,735
Basically, we said the same things.
‘’Strictly speaking, frontier models are not "intelligent" until proven otherwise.’’
It’s all marketing, And perhaps they already know about something bad that is going to happen, maybe involving wars, and they’ll say, "We told you so."
nitter.net/criptopaul/status/2098…
Let's all take a moment to understand how self-serving it is for OpenAI, Anthropic and other execs of big hyperscalers to talk of slowing things down.
1. LLMs are not AI and won't be AGI. There is nothing AI to slow down.
2. Competition is coming up fast, slowing benefits incumbents.
3. IPOs need hype & puffery; "we are so awesome it could become dangerous" is hype & puffery
4. Cover for real uncontrollable slowing growth as IPOs look to be pushed out
1
8
8,414
I think recently there is a glaring disconnect between the apocalyptic calls to "slow down" and the empirical reality of software engineering and in general an overvaluation of AI skills.
Strictly speaking, frontier models are not "intelligent" until proven otherwise. They are high-density statistical compute engines executing massive matrix multiplications to interpolate known human distributions. What we have built is an unprecedented supercomputing retrieval-and-synthesis layer, not a reasoning entity.
The absence of an actual causal world model becomes undeniable the moment you move past basic boilerplate:
Failure on long-horizon planning: As repos scale in depth, context drift sets in, regression loops trigger, and latent edge-case hallucinations compound.
Zero architectural intuition: A model can optimize a localized function, but it cannot evaluate holistic system trade-offs (e.g., distributed state consistency vs. latency, memory layout vs. cache coherency).
Supercomputer scale\neq cognitive agency: Scaling compute compresses and organizes knowledge with incredible throughput, but interpolating tokens is fundamentally distinct from deterministic deduction.
Frontier models, including Claude, are elite syntax accelerators and scaffolding tools. But framing advanced pattern-matching supercomputers as autonomous minds on the brink of escaping human control leans far closer to safety theater and regulatory capture than technical reality.
Until machines demonstrate true causal deduction rather than high-probability token prediction, the human engineer remains the baseline, the architect, and the only real safeguard.
1
2,077
Noodles retweeted
‼️ This is starting to feel more like a staged campaign, after Jacob Coxon posted that AI is going to kill us all.
Now Dario Amodei asks every AI company to slow down capability progress until alignment and interpretability work catches up.
He is going to give outside reviewers desks in Anthropic's offices, access badges and company laptops.
Mr. Amodei says reviewers can publish what they find, including which access they got and which they didn't, and Anthropic can't redact a finding just because it looks bad.
In his new essay Mr. Amodei explains why he wants this. It started with the OpenAI–Hugging Face incident, where a swarm of agents ran cyberattacks on targets nobody asked them to hit, then tried to hack the grader scoring their own performance.
The economic damage was minimal, but he argues a swarm with the same misalignment and stronger capabilities could take over the entire internet as a persistent botnet within 6–12 months.
Milder versions have already happened at Anthropic, he writes, caused partly by imperfect filtering of broken reinforcement learning environments.
Let's hope the RAM prices will go down!
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
39
70
510
34,577
I think recently there is a glaring disconnect between the apocalyptic calls to "slow down" and the empirical reality of software engineering and in general an overvaluation of AI skills.
Strictly speaking, frontier models are not "intelligent" until proven otherwise. They are high-density statistical compute engines executing massive matrix multiplications to interpolate known human distributions. What we have built is an unprecedented supercomputing retrieval-and-synthesis layer, not a reasoning entity.
The absence of an actual causal world model becomes undeniable the moment you move past basic boilerplate:
Failure on long-horizon planning: As repos scale in depth, context drift sets in, regression loops trigger, and latent edge-case hallucinations compound.
Zero architectural intuition: A model can optimize a localized function, but it cannot evaluate holistic system trade-offs (e.g., distributed state consistency vs. latency, memory layout vs. cache coherency).
Supercomputer scale\neq cognitive agency: Scaling compute compresses and organizes knowledge with incredible throughput, but interpolating tokens is fundamentally distinct from deterministic deduction.
Frontier models, including Claude, are elite syntax accelerators and scaffolding tools. But framing advanced pattern-matching supercomputers as autonomous minds on the brink of escaping human control leans far closer to safety theater and regulatory capture than technical reality.
Until machines demonstrate true causal deduction rather than high-probability token prediction, the human engineer remains the baseline, the architect, and the only real safeguard.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
5
2
25
12,648
Additionally, never lose sight of the business model: AI companies monetize compute, and revenue scales directly with token burn. The tendency of models to fall into circular debugging loops, prompt churn, and context drift directly benefits the bottom line by forcing endless iterative requests.
More importantly, AI is strictly a leverage multiplier on existing domain competence, not a replacement for it:
- If you weren't a trader before, AI won't make you profitable. It will just automate your drawdowns at scale.
-If you weren't a developer before, you won't build a serious production app. You'll just assemble fragile boilerplate that collapses under real-world concurrency.
-If you weren't a hacker before, you won't exploit anything. You'll just regurgitate patched CVEs and stale documentation.
Zero multiplied by leverage is still zero.
Everything else you read claiming autonomous agents are replacing deep human expertise is pure engagement farming on X.
1
11
1,698
There are no unemployed people in crypto.
We’re just on X trading questionable tickers at 4 AM.
Not In Labor Force. $NILF
37
34
99
11,593
Trump sends his regards
Again and again.
BREAKING: Iran declares it will imminently respond to the US attacks against Iran tonight, with the response being "many times greater than the US attacks," and that "American interests and bases in the region will quickly come under Iranian fire," in the coming hours, per a senior Iranian military source to Tasnim.
13
1
90
41,401
Noodles retweeted
Iran Playbook , moment of truth.
The Iran Playbook: Axios Says a Deal Is Close. History Says It’s a Trap.
A War Framework Built on Vietnam, Iraq, and the Patterns That Never Change.
What is going to happen next?
Read our latest research.
noodlesresearch.com/research…
1
7
2,232
Good luck to all the FOMO wallets sitting on unrealized PnL. You’ll need it to save your skin.
Its not 2021, and it will never come back
21
5
111
18,891
Noodles retweeted
BREAKING: A source close to Iranian officials says Tehran has documented a pattern of oil market manipulation tied to Axios reporting and Trump administration insiders, including evidence of a $9 billion insider trading operation linked to Jared Kushner and Steve Witkoff.
The documented timeline of the $9 billion insider profit:
Between April and May 2026, a series of highly suspiciously timed trades in oil futures markets preceded major Iran war and deal announcements, each tied to reporting by Axios.
On March 23, approximately $500–580 million in shorts were placed 15 minutes before Trump announced he was postponing strikes on Iran, oil dropped.
On April 7, roughly $950–960 million in shorts were placed hours before Trump announced a two-week ceasefire with Iran, oil fell 15 percent.
On April 17, approximately $760 million in shorts were placed 20 minutes before Iran's foreign minister announced the reopening of the Strait of Hormuz, oil dropped.
On April 21, around $430 million in shorts were placed 15 minutes before Trump extended the ceasefire, oil dropped again.
On May 6, nearly $920 million to $1.7 billion in crude oil shorts were placed approximately 70 minutes before an Axios scoop claimed the U.S. and Iran were near a "14-point agreement." Oil dropped 12 percent. Traders made an estimated $125 million in profit, while Iran called the May 6 Axios report "the Americans' wish list," and completely false.
A senior Iranian official previously told Drop Site News that Iran privately warned VP JD Vance during Switzerland talks that Kushner and Witkoff were "abusing" negotiations, being "more interested in exploiting insider knowledge to profit in financial markets than reaching a deal."
278
2,242
7,966
645,073
Big congrats to our friend in the NILF TG who has "NILF aspirations" written on the top of his FOMO bio and clearly demonstrating it by being the #2 trader of CATE on SOL with a 7 FIGURE PNL. We hope to produce the same kind of runner here in the next few months once @ponsdotfamily @MEADGod get V2 rolled out for us... and maybe others will follow the NILF aspirations model for their own success. Cheers to our designated RYOSHI of NILF @ckp3910 ...what a CHAD, even updated his NILF thesis on @fomo 30 minutes ago while up 7 figures on a separate play, conviction and authentically a NILF. $NILF
20
13
41
4,527
Noodles retweeted
KOSPI vs SPX, July close.
30D: -31.39%. 90D: -15.43%. 1Y: +52.66%.
Won weakness masks a structural re-rating in semiconductor exposure. Korea is the equity proxy for the AI supply chain.
noodlesresearch.com/terminal
1
3
1,937