The most important thing in markets is the rate of change. This page tracks how the
conviction of the experts we follow moved, topic by topic, versus a recent baseline.
A synthesis of what the experts said recently — not a Synthos market call.
For each of 33 tracked topics we measure the net stance of the voices we track over the last 4 weeks,
then compare it to the prior 8-week baseline. The gap is the delta — the rate of change
that actually matters. No grades, no scorekeeping.
The last 4 weeks’ conversation: 2081 attributed claims
— 1021 bullish ·
619 neutral ·
441 bearish.
6 of 33 topics moved past the change threshold this cycle;
22 held, and 5 had too little signal to call.
This month at a glance
6 topics moved past the change threshold — 3 turned more bullish, 3 more bearish vs. the prior baseline. Each bar shows the size and direction of the shift; full detail below.
Copper / industrial metals+30 points · turned more bullish
Space / SpaceX+23 points · turned more bullish
Oil & energy+15 points · turned more bullish
Inflation vs. deflation−16 points · turned more bearish
US dollar−19 points · turned more bearish
Frontier AI labs−34 points · turned more bearish
The Delta — what moved
Topics whose conviction shifted by a meaningful, reliable margin over the last 4 weeks
versus the prior baseline: a change of at least 15 points on our −100…+100 scale, on at least
25 claims, in a window that wasn’t mostly neutral, and confirmed by a real shift in the
bull-versus-bear balance (not just quieter talk). Largest move first. Each shows the sentiment label before and after, the size and direction
in words, the window’s volume and coverage, and the sourced claims driving it.
Frontier AI labs
AI stack
Neutral / Mixed → changed to Neutral / Mixed
Still Neutral / Mixed as a label, but the underlying conviction moved −34 points.
−34 points · turned more bearish32 claims in windowFull-coverage reading
Why it moved: What drove it
“Their whole thesis: models won't eat the whole stack; the world fractures into specialized workflows and proprietary data, with model routers as the general-purpose layer.”
“Kimi K3 is profitable at $3/$15 per million tokens (~Sonnet price); similarly-sized closed models charging $10/$50 must have mindboggling margins — selling tokens is an incredible business, maybe better than…”
“Gemini is clearly in fifth place and may stay there maybe forever; Google signed long-term compute deals it can't claw back, signaling leadership lacks conviction it can reach RSI.”
“Ability to charge a premium for frontier models will disappear in most industries within 12-18 months as cheaper open-source models improve and most workflows don't need the frontier.”
“Copper is the leading metal at the year's highs; China warehouse inventories drawn more than expected plus production down at most major Chilean mines gives a good short-to-medium-term supply-demand dynamic.”
+23 points · turned more bullish26 claims in windowFull-coverage reading
Why it moved: What drove it
“SpaceX is Baron's largest position (~$25B of $70B); expects 20-30x from IPO price over next 10-15 years and a $20-30T (at least $40T) valuation — biggest company in the world.”
“SpaceX is Baron's biggest position (~$25B of $70B); will be the biggest company in the world, worth $20-30T, at least $40T; make 20-30x from the IPO price.”
“At a $1.4T valuation (down 40-50% from IPO) SpaceX is compelling; guides are extraordinary — 100B ARR by year-end and $1T ARR pulled forward to 2030 — and the market isn't pricing it in. Holding for grandkids.”
“The dollar has structurally declined in purchasing power for multiple decades—priced in units of dollars, the S&P, CPI and real goods all show the dollar falling since Nixon ended gold convertibility in 1971.”
“Bessant sold ESF euros and leaned on the FIMA facility to buy yen without selling bonds — engineering a lower DXY without spooking the bond market, which is liquidity-positive.”
“June core CPI fell fractionally (first since 2020) and TIPS break-evens are plummeting; the energy shock brings demand destruction, not second-round inflation — inflation risk as close to zero as a complex…”
“TIPS breakevens and the flat yield curve say unequivocally there's no inflation risk beyond the short run; CPI fell the most since 2020 and PPI came in far weaker with no second-round effects from oil.”
“30-year yield rose 4.02%→5.27%, up 125bps during Fed cuts — biggest in 40 years; market is signaling inflation; if the Fed cuts now the 30-year heads to 6%.”
+15 points · turned more bullish103 claims in windowFull-coverage reading
Why it moved: What drove it
“Oil is an input, not a variable to forecast — it sets the pace of Iran-US conflict; it stays range-bound ~80-100, and spikes toward 100-120 force both sides to de-escalate, then it falls back.”
“There's a glut of crude oil; the $125 wartime peak was the top and the big spike everyone predicts is not coming — the market is telling us prices head lower.”
“After 18 months of well-level study, US faces historic structural gas deficit from LNG exports (15→35 BCF/d) plus AI compute; upside price risk is convex and unbounded from ~$3.50 today.”
“Expand Energy is far and away the biggest winner — controls ~70% of remaining core Haynesville wells, highest-quality rock, yet trades ~4x forward EBITDA after CEO-search-driven selloff.”
Topics with enough claims to read reliably that held over the window — no move past
the change threshold — grouped by theme. “Drift” is the small, sub-threshold direction of
travel versus baseline, in words.
Macro & rates
Macro & rates: topics that held over the window, with current sentiment label, claim volume, and sub-threshold drift
Topic
Current sentiment
Claims
Drift vs. baseline
Fed & interest rates
Neutral / Mixed
102
slightly more bullish
Fiscal dominance / debasement
Very Bullish
54
slightly more bearish
Labor market
Very Bearish
23
slightly more bullish
Liquidity & financial conditions
Bullish
35
slightly more bullish
Private credit / credit stress
Very Bearish
30
flat
Equities & structure
Equities & structure: topics that held over the window, with current sentiment label, claim volume, and sub-threshold drift
Topic
Current sentiment
Claims
Drift vs. baseline
Market structure / passive flows
Neutral / Mixed
39
slightly more bullish
US equities / S&P 500
Bullish
120
slightly more bullish
AI stack
AI stack: topics that held over the window, with current sentiment label, claim volume, and sub-threshold drift
Topic
Current sentiment
Claims
Drift vs. baseline
AI capex / buildout
Bullish
310
flat
AI software & agents
Very Bullish
49
slightly more bullish
Datacenter power
Very Bullish
32
slightly more bearish
Memory / HBM
Bullish
58
slightly more bullish
Semiconductors / GPUs
Bullish
87
slightly more bearish
Crypto
Crypto: topics that held over the window, with current sentiment label, claim volume, and sub-threshold drift
Topic
Current sentiment
Claims
Drift vs. baseline
Bitcoin
Bullish
149
slightly more bearish
Ethereum & L1s
Bullish
41
slightly more bearish
On-chain finance / crypto infra
Very Bullish
42
slightly more bearish
Stablecoins & tokenization
Very Bullish
23
slightly more bullish
Commodities & hard assets
Commodities & hard assets: topics that held over the window, with current sentiment label, claim volume, and sub-threshold drift
Topic
Current sentiment
Claims
Drift vs. baseline
Gold
Very Bullish
108
slightly more bullish
Silver & precious metals
Very Bullish
48
slightly more bullish
Geopolitics & international
Geopolitics & international: topics that held over the window, with current sentiment label, claim volume, and sub-threshold drift
Topic
Current sentiment
Claims
Drift vs. baseline
Geopolitical conflict risk
Neutral / Mixed
32
flat
International / ex-US markets
Neutral / Mixed
11
slightly more bullish
Sectors & exponential themes
Sectors & exponential themes: topics that held over the window, with current sentiment label, claim volume, and sub-threshold drift
Topic
Current sentiment
Claims
Drift vs. baseline
GLP-1 / obesity / Eli Lilly
Very Bullish
18
flat
NVIDIA
Bullish
33
slightly more bearish
Too thin to call
Tracked topics where we’re deliberately not reporting a move — too few
claims in the window, a mostly-neutral read with no clear direction, or a swing the bull/bear balance didn’t
confirm. Showing this rather than a flashy but unreliable number is the point: a −80 on ten claims is
noise, not news.
Topics with too little signal in the window to report a change, with current sentiment label, claim volume, and the reason