The Return Of Inflation
Last week’s inflation print supported the value of midstream energy in a portfolio for preservation of purchasing power. We’ve noted this recently (see A Healthy Correction last paragraph). The most recent 3.8% CPI reading is driving longer term inflation expectations higher.
Incoming Fed chair Kevin Warsh has his work cut out for him satisfying both President Trump and the bond market. On Wednesday, the US Treasury auctioned a 30 year bond with a 5% yield for the first time in almost two decades. The market knows Trump is not a hard money man.
Last week Spiro Dounis, who covers midstream energy for Citigroup, raised his price target for Venture Global and initiated coverage on NextDecade. Both stocks were firm last week – the fundamental justified it, but perhaps Citi’s improved outlook helped. We think the market continues to underestimate the benefit to US LNG exporters of the supply disruption from Qatar.
I have a few thoughts on AI.
Nobody need look at the web results from a Google search anymore – just go with their AI model Gemini. Or skip Google entirely with Claude, ChatGPT or Grok. Ask incredibly detailed, obscure questions and marvel at the results.
I had a printer problem recently – I interacted via chat with an HP AI agent that asked sensible questions and had me check Windows processes I’d never seen before. Eventually it concluded the problem was too hard and referred me to a human agent. I was advised to use simple words so that the instant translation to their presumably non-English speaking human would be error free. The printer was soon working again.
Clearly, AI is impacting the IT industry. Meta is laying off 10% of its staff, because AI can do their jobs but also to pay for their vast investments in data centers. Programming is just the kind of rules-based activity that lends itself to AI. Accounting is another. Maybe the list of jobs not impacted is the shorter one. It’s already disrupting some industries, and the sums being invested suggest profound change everywhere.
But it’s not there yet and given the hype and money deployed I’m starting to wonder why it’s not better than it is. For example, I uploaded a form to ChatGPT that we had to prepare for a business partner, along with last year’s version, asking it to complete as much of the new one as possible using responses from the old one. It produced a form with questions out of sequence, questions inserted as answers, and questions deleted. It was as if a newly hired distracted admin had done it quickly while focused on something else like TikTok. There was no proof-reading – AI thinks none is needed because it’s never wrong.
Someone I know asked Microsoft’s Co-Pilot to re-engineer a complex spreadsheet so that it would run faster. Its first effort deleted all the data, which did speed things up. AI doesn’t have common sense, but does have supreme self-confidence in its results.
I’ve debated with people whether AI will eliminate financial advisors, or doctors. I know the impact will be profound. But I also think a 55 year old thinking about retirement will want to talk to a human advisor when the market’s down 10%. And a cancer patient will want a human oncologist to guide her treatment. You’ll want your professional to use AI as a tool to help but not be replaced by it.
Humans have accountability; software does not.
Asking your preferred AI model detailed questions is free or nearly so. I expect this is to get us hooked so that life can’t be lived without the answer to every possible question being available on your phone. And then prices will rise.
Owning a device with access to all known facts is empowering. But I haven’t found that AI can do any actual work. Processing all the world’s written knowledge instills hubris but not common sense. As we use it more, we’ll get more dumb answers and nonsensical documents and spreadsheets.
We can generally tell when it’s garbage. But we’re heading in a direction where the questions will be too complex for us to check that the answers are reasonable or even verify them at all. It needs to get better fast, because eventually we’ll have to have faith in AI. If by then it hasn’t earned our trust with sensible responses to the simple questions, it won’t make that next leap.
An SNL skit poking fun at AI will be a sign that the world’s not changing quite so fast.
We have two have funds that seek to profit from this environment: