Blain’s Morning Porridge – Jan 12th 2023: ChatGPT vs Bill Blain on Alternatives, and how it’s going to change the world
“They have a name for the winners in the world, I want a name when I loose…”
This morning: Microsoft taking a controlling stake in ChatGPT highlights how the world is going to change as AI becomes embedded across news, data and analysis. Compare my take on the market with the AI. As GPT rolls out it will trigger a new Tech revolution, but how will it be controlled and what are the dangers?
Two things to report before I start this morning:
First – Any Steely Dan fans out there? Last night I went to a superb gig in Ronnie Scott’s, the premier London Jazz Club. The Steely Dan tribute band was superb – The Royal Scammers are so professional they have Dave Arch, the musical director of the BBC’s Strictly Come Dancing, as the Piano player! I went with my daughter – she grew up to the sounds of The Dan and loved it. I saw Steely Dan a couple of times back in the day, and this band were just as good! (And that is really saying something…) My only query: why didn’t they play Dr Wu?
Second – Somewhat embarrassingly, I’m delighted (but concerned in case I come over as an idiot) to be featured in the excellent London Financial daily City AM this morning: The Square Mile and Me: Bill Blain on Bear Stearns, the Old Dr Butler’s Head and sailing on the Solent. I’d like to know what people really think… Don’t hold back.
Back to the unreal world…
ChatGPT – Inklings of a very different future
I probably should be thinking about today’s US inflation data, but since this is the real world, let’s take the blue pill and dive into the future of Artificial Intelligence – and scare ourselves about the future.
Big news yesterday was Microsoft looking to invest $10 billion into OpenAI’s language generation model, ChatGPT – effectively taking a majority stake in the company now valued at $29 bln. It makes sense for the Tech giant to expand into an AI vision that could potentially overturn the current tech consensus through language processing. It will totally redefine the way we search, digest and use information on the internet – changing our decision making processes in multiple ways, and influencing investment consensus if we’re ultimately all driven by the same AI thinking. That’s something any Tech firm wants to be at the centre of.
Many analysts believe ChatCPT and other “virtual assistants” are going to become Google Killers – reinventing the search engine. It will do far more. In the World of Big Tech, the only way to survive is through constant reinvention and redefinition of what you are. For financial bloggers and strategists – like myself… its got enormous implications.
I’ve been wondering just how quickly the programme will make me redundant. It’s a theme my young Gen-Zed colleague, Daren Revan, who is Shard’s media guru, keeps reminding me of: how Pale, Male and Stale ancient has-beens like myself will be replaced by smarter computer chips…. It’s got me worried.
So we decided to do a test.
Daren posed the question: “What are some interesting alternative asset themes for 2023?” and then gave me and the chatbot 5 mins to respond.
My immediate response was:
Alternative Assets are generally off-market, private, illiquid, non-transparent and require a high degree of sophistication and bilateral price-setting. They are likely to remain one of the most interesting investment themes in 2023, driven by investors looking to hedge their investments against the ongoing uncertainty in stocks and bonds with predictable income streams. [See, I know my onions…]
Markets are struggling to determine how inflation, interest rates, central bank normalisation, and the recession/stagflation threat develops. The risks are we see further rises in bond yields and an ongoing correction to equities as global markets adjust to the end of easy money since 2009 and the normalisation of the post QE era. I expect alternative non-mainstream assets will be a prime focus for wealth managers looking to generate stable, predictable off-market returns in what is an increasingly difficult to predict world.
The trick to Alternatives is deep-dive due diligence – choosing investments most likely to generate the returns they promise in the pitch. That means being sure of issues such as the technology, the feedstock agreements, the offtake sales, potential subsidies, security over assets, the management’s skill set, and to test all the assumptions about income to make sure it will generate the promised cashflows. It requires knowledge and insight to get right. Its clever. It’s smart. It’s based on insights and knowledge.
Examples of how investors should be approaching alternatives would include:
- Aircraft leasing – will the rentals on the airplane be achieved and is the airline renting the plane likely to survive any downside scenarios? (This applies to all leased and rented assets – any asset with associated cashflows.)
- New Technology – what is the need, the potential market, and the upside/downside risks.
- Private Equity – in a recessionary environment how many private equity investments will survive high interest rates on their debt, and more realistic earnings assumptions?
- Private debt – How will higher rates impact smaller companies using debt to finance plant and costs – and how are corporate defaults set to rise?
- Energy Transition – What are real benefits from renewable energy investments across the whole energy life cycle – wind is cheap but unreliable, tidal is expensive but completely predictable. How much more attention will markets pay to real carbon costs…
I waffled on like this for a couple of more paragraphs – but broadly my message is Alternatives are great, but only if you really understand the risks and opportunities.
The Chatbot was more focused and answered the question literally– this is how it genuinely responded to the same question:
“Some alternative asset themes that may be interesting for 2023 include:
- Infrastructure: Investment in infrastructure assets, such as transportation, energy, and telecommunications, is expected to continue to grow as governments seek to improve their infrastructure and stimulate economic growth.
- Real Estate: Real estate investment trusts (REITs) and other real estate-related investments may be attractive as the economy recovers and interest rates remain low.
- Private Equity: Private equity funds and venture capital firms may provide opportunities for investors to invest in companies that are not publicly traded, but have strong growth potential.
- Environmental, Social, and Governance (ESG): As investors become increasingly aware of the importance of environmental, social, and governance factors, investments that focus on these areas may be more in demand.
- Cryptocurrency: Cryptocurrency and blockchain technology continue to gain mainstream acceptance, so it could be a promising theme of investment for the next year.
- Technology Advancements: Advancements in fields such as biotech, AI and 5G technology are expected to continue, and investments in these fields may provide opportunities for growth.
The Chatbot took seconds to give that literal answer to the question it was asked. I delved into the topic – thinking (slowly) about it. I really can’t fault the programme – except to say I can already get the same answers from any typical investment bank when it comes to making bland but somewhat interesting statements.
However, I can definitely see why applying the chatbot to any financial question looks like an interesting way to start probing the question.
I don’t mind admitting I find ChatGPT more than a little scary.
It’s a whole new way of using data and info. It’s going to totally overturn the current way we share and use everything from news, data, academia, and literature. I fear original thought will be buried in mass produced AI verbiage. The risk is we lose our own critical faculties if we rely on AI to do the thinking – the ultimate Sci-Fi horror.
How surprised was I to discover OpenAI has Elon Musk as a founder and investor? – alongside a host of other tech luminaries. Musk resigned from the board over conflict of interest with Tesla’s own AI. He remains invested.
OpenAI has morphed from a AI Not-for-profit lab pioneering “friendly AI” into a full-on business. I used the OpenAI programme to discover that ChatGPT stands for “Conversational Generative Pre-training Transformer.” It says it is: “a cutting-edge language model that uses deep learning techniques to generate human-like text. The model has been trained on a massive dataset of conversational text, allowing it to understand context and generate responses that are highly coherent and engaging.”
If that’s true, I am in trouble. So I will simply write; “The cherry trees are blooming on the Road to Smolensk”, and hope the chatbot struggles to interrogate that statement. “Niet Comrade, the Snow falls heavily on Kharkov Avenue.” (I figure if I write bollox like this, then the chatbot might give up?)
Already I’m being deluged by messages on Linked-in offering to show me CVs for similar “virtual assistants” I can “hire” to take the pressure off, and invitations to use AI on the Morning Porridge. Of course I never will.. I will switch off the Hal900 before I ever allow that … “I’m sorry Dave, I can’t allow you to do that.. ….. Daisy, Daisy… give me your answer do….” (People may be struggling with some of the references this morning but check out The Cherry Orchard and Space Oddessy.)
I can imagine there is a whole generation of young bankers, currently writing deep and meaningful deal pitchbooks who are going to find themselves out of a job tomorrow. But, writing pitchbooks is a skill builder. It trains bankers to think logically about how to present deals and investments, and how to counter buyer concerns. If the computer does it – who is learning?
I encountered Microsoft in the 1980s as I experimented with the first spreadsheets and word processors. I built my first chart analysis on Lotus Notes before it was absorbed by Excel. By the 1990s MS was getting annoying with that paperclip thing you just wanted to punch. What this new iterative AI will offer us is debatable, but you can guarantee Microsoft will be deluging us with virtual assistants across its Office365 and Outlook platforms to plan our days and email responses, chatbots to answer internet searches – even before you think of what you want to ask, instant translation services, and a host of other stuff we never even though of – yet.
Some people might think that sounds exciting. Maybe it will make us more productive ad improve business.
To me it sounds like a dystopian nightmare..
Sooner that we expect, these new chatbot/GPT AIs will have wormed into our lives and will be operating right across the Microsoft product suite, inserting themselves into every aspect of our lives and the programme we use.
I will need to go off grid and start posting the Morning porridge on Papyrus. I don’t want a “human-like conversational experience.” I want to chat to my clients and write stuff that interests them.
Five Things To Read This Morning
Guardian College student claims app ca detect essays written by Chatbot ChatGPT
FT Goldman Sachs makes brutal job cuts in quest for lower costs
Bberg Homebuilder Persimmon Offers Mortgage Holidays to Lure Reluctant UK Buyers
Torygraph More than 750,000 UK households at risk of mortgage default
WSJ JPMorgan Bought College Financial-Aid Platform and most of the users were fake!
Plus
Out of time, and back to the day job… and still coughing with this damn flu…
Bill Blain
Shard Capital
6 Comments
Comments are closed.
I really enjoyed the write-up in City AM. I didn’t know you were ex-Euromoney. So am I, but on the legal and tax journalism side of things and a few years after you. Padraic Fallon was a legend. All the best,
“A random walk down wall street” or as more commonly known “where are all the customers yachts?” hints at the issue/problem with what is being called AI these days. It is, albeit very cleverly constructed and sophisticated, nothing more than regression and much like technicial analysis its problem lies in the fact that if the past accurately forecast the future……. well life would be pretty boring and society would probably have a fuedal architecture (bring back the dark ages?). So combining all the information on the net with filters for the more obvious sinks will seem novel for a little while. I wouldnt worry about the day job as nothing io have seen from ChatGPT can touch the morning porridge. Sure people whose job is aggregating and synthesing existing info will suffer but the thing that drives humans is novelty, surprise and curiosity and you already have spotted that when you describe the output as bland. Without expertise it can be hard to judge but thats the signpost I would suggest.
Don’t worry Bill. The more correct but obvious generalist content will be posted, the more we need/ like real shocking original research. The Few people who still have an opinion which challenge our own thoughts will be the real heroes So keep up the good work and you will be rewarded
As I have discovered, Chat Boit is incapable of either sarcasm or conscious irony. Plus – very weak on metaphor or most kinds of subtlety.
Hmmm, I am excited about the research possibilities using A.I.
Investing, global affairs, curing cancer, ending the war in Ukraine – a very deep well indeed
Mostly I view A.I. as a tremendous idea generator but it will still take a human, like our good Mr. Blain, to make sense of it all and put a personal stamp on it.
Fear nothing.
great comment from a student (my nephew, Tom Blain, reading fiendishly complex Maths at Bristol):
“If you want to keep your job just make sure you are open to embrace AI and learn how to work with it, rather than against it. Understand the capabilities and limitations of the model to utilise it’s power for your own skillset.
With ChatGPT, maybe you can think of it as someone who has skimmed almost every scientific paper and article ever published, but then fell into a deep coma at some point in 2021 – when the training data cuts off. If you ask it, “what is going to happen in 2023”, it can only make a very generic guess. It has no knowledge of any current world events and does not generate new data, its only a language model.
I’ve found it useful when you ask a more specific questions: “why is … a drawback of … in this scenario … and what can we do to prevent …”.
Or especially useful when I ask it to write my essays the night before deadline when i’m at the pub.”