Inflation is elevated (~2.93% core), not near Fed target, and will take time. Real economy question is labor market. Job creation has slowed dramatically, but needed job growth may be only ~25k/month due to collapsed immigration (CBO estimate fell from 2.7M to 48k). This means weak job numbers might be 'okay' given new demographic reality. Risk is Fed cuts too aggressively, re-sparking inflation.

Population growth is at 80-year low due to border closure/deportations. J. Powell suggested breakeven job growth could be 0-50k.

AI data center buildout adds 1.25-1.5% to GDP now ($1T+ capex). Long-term, AI will follow Bill Gates' internet rule: underdeliver 1-3 years, overdeliver 10 years. Immediate productivity gain is in coding. Long-run, tech advances are net job creators, but new jobs are in industries that don't exist yet (like internet created).

Cites Northwestern professor Bob Gordon: major tech advances are net job creators in long run. Example: cheaper transportation from autonomous vehicles will enable new business models.

AI will become a deflationary pull several years from now as it becomes ubiquitous. Current sticky inflation (~3%) is due to post-COVID structural changes: remote work (up from 7% to 27% of workforce), deglobalization/segmentation, and reversed immigration patterns globally. These offset prior tech efficiencies.

Remote work shifts power to workers, sustaining ~3% wage growth. If 3% is inflation baseline, Fed might be done cutting rates.

Defends Moore for providing heterodox views, breaking Fed groupthink. Criticizes NY Fed's John Williams for singling out Moore. Highlights Fed's lack of accountability, citing 2019 op-ed by Bill Dudley suggesting Fed could tank economy to affect election, and Lael Brainard's 2016 donations to Clinton while arguing against rate hikes.

Fed's unanimous votes provide no information, forcing reliance on dot plot. Bank of England's split votes provide better forward guidance.

Independence is a continuum. Risk is pendulum swings too far: from Brainard's alleged conflict to Trump's desire for aggressive cuts to lower debt costs (fiscal dominance). If Fed cuts to 1% for fiscal reasons, it could never raise again without triggering market vigilantes.

Congress, not Fed, is responsible for debt burden. Fed's job is not to manage Treasury's interest costs.

Agrees managed rates reduce informational value. Yield curve was perfect pre-COVID indicator but hasn't predicted post-2020 recession due to being 'overly managed.' Historically, 2s10s spread is ~100bps. Argues short-term rates (~4%) are about right (neutral-ish), but long-term rates are too low and should go up.

In 3% inflation world, with neutral rate ~1% above inflation, neutral funds rate is 3.5-4%. We're at 4% now.

Index is short duration (90% of benchmark), bulleted in 5-10y sector (curve steepener), underweight corporates (spreads at generational tights), overweight mortgages, with conviction bets in local EM debt (6.8% yield) and short-term TIPS. Expects long-term rates to rise, short-term rates held steady by Fed cuts, curve to steepen.

Corporate bond spreads correlate with stock market valuations; prefers mortgage risk. Index outperformed Agg by ~100bps in year ending Sept.

Maintains 4% cash, 5% bonds, 6% stocks forward return expectation. Stock market's 14% YTD gain is 70% from 41 AI stocks; other 459 stocks are near 6%. AI stocks are 47% of S&P, merging S&P and NDX composition. High valuations mean market likely disappoints over medium term, averaging 6%.

Valuation is not a timing tool but sets expectations. Returns could average 6% via a crash then recovery or sustained low single-digit growth.

AI bubble parallels 2000 internet: not that valuations are universally insane (some are), but market prices every AI stock as a winner. Not all can win. When bubble pops, losers drop 80-90%, winners don't rise much further.

Cites warnings from Sam Altman, Jeff Bezos, David Solomon. In 2000, any '.com' name caused stock spikes.