A daily market update from FS Insight — what you need to know ahead of opening bell

“The person across the table is never the problem. The unsolved issue is. So focus on the issue.” — Christopher Voss

Overnight

Oil futures settle at fresh 7-week highs as U.S. crude and gasoline supplies decline (MS)

Nvidia, the Nasdaq, and the S&P500 got dizzy at record highs on Thursday as world markets start to take stock of a bumper 2024 (RT)

U.S. labor market, housing data point to slowing economy (RT)

BoE held rates steady but readies to cut soon (WSJ)

Citi bets on Europe despite political uncertainty (RT)

Sycamore CEO says persistent high rates are boosting public markets (BBG)

Apollo’s chief economist expects no rate cuts this year (RT)

Supreme Court sidesteps wealth-tax question in closely watched case (FT)

Swiss Central Bank cuts rate for second straight time (Barron’s)

Iran’s reformist candidate jolts presidential race with pitch to restart nuclear talks, entertain closer relations with the West, adopt softer stance on the hijab (FT)

Banks pressed to screen for money flow from Fentanyl precursors (WSJ)

Accenturerose ~7% after forecasting annual revenue growth above expectations on surging demand for AI integration services (RT)

IRS to reject billions of dollars in Covid employer tax-credit claims (WSJ)

Most U.S. car dealerships still hit with massive computer system outage (CNN)

Paris gets a wild meme stock as Atos becomes a chatroom favorite (BBG)

Carlyle expands energy assets with Energean portfolio buy of up to $945 million (WSJ)

Carlyle and KKR vie for Discover’s $10bn portfolio of US student loans (FT)

British fintech Revolut seeks valuation of more than $40bn in employee share sale (FT)

Two people arrested in London on suspicion of running illegal cryptoasset exchange (IW)

Hedge fund talent schools are looking for the perfect trader (YH)

Investors say Labour could borrow more without U.K. bond market backlash (FT)

Anthropic debuts its latest ‘industry-beating’ AI model (TC)

Deutsche Bank’s private bank cuts consultant spending by 70% (FT)

Large U.S. banks face growing spillover risks from non-banks (RT)

U.S. finds no dollar manipulation in 2023, but adds Japan to monitoring list (RT)

Green bond issuance surges as investors hunt for yield (FT)

Taiwan’s central bank warns of systemic risks of rapid ETF growth (FT)

SEC urges U.K. to pick date for stock market shake-up (RT)

KPMG will cut another 200 U.K. jobs amid slowdown (FT)

PayPal names Walmart exec Venkatesan as new CTO (RT)

Booking’s CEO slams E.U. over ‘dumb’ regulations (FT)

Birth rates in rich countries halve to hit record low (FT)

Donald Sutherland, versatile star of ‘M*A*S*H,’ ‘Ordinary People’ and ‘The Hunger Games,’ dies at 88 (HR)

Chart of the Day

MARKET LEVELS

MORNING INSIGHT

Good morning!

The S&P 500 has gained +0.75% in this shortened week and is tracking to be up ~4% for the month. Equities have risen in June for the ‘right reasons’, in our view, as inflation has been tracking softer than consensus. The most important incremental inflation news this week has been in autos.

New car inventory has surged back to pre-pandemic levels as auto production has improved and consumer demand has waned. In fact, Ford now has 100 days’ supply of cars. These are high levels, and this is not just Ford. As we highlighted last week, new-car incentives have surged.

Click HERE for more.

TECHNICAL

Thursday brought some minor rotation out of Technology, and given the recent run-up in QQQ relative to SPX, as well as strong outperformance by Technology in the last month, it looks likely that some backing and filling might be possible.

Both NVDA and AAPL now show daily counter-trend exhaustion per DeMark’s TD Sequential indicators; however, it won’t be possible to confirm weekly exhaustion in either until next week.

As discussed in recent days, the degree to which Technology has begun to outperform all parts of the U.S. stock market has grown more pronounced. Yet this doesn’t translate into an imminent need for a big selloff – for the broader market.

In recent days, U.S. Equities have shown some improvement in Financials, Healthcare, Industrials, and Energy which could lead to a short-term bounce in the Equal-weighted S&P following some severe underperformance since March, compared to the SPX.

Click HERE for more.

CRYPTO

Cross-chain infrastructure protocol LayerZero opened claims for its native token, ZRO 0.74% , this morning. LayerZero implemented a new caveat to airdrop claims, having users contribute $0.10 per ZRO to Protocol Guild in order to claim. LayerZero Foundation estimates the mechanism will raise $18.5 million for the Ethereum development non-profit, and LayerZero said they will match up to $10 million in donations. The “proof-of-donation” mechanism has not been well met by the community, with some taking the stance that airdrops should be a reward for interacting with the protocol and not a pay-to-claim function. LayerZero previously completed a Series B funding round in April 2023, valuing the company at $3 billion. ZRO is currently trading at approximately $3.45, implying an FDV of $3.45 billion, meaning investors in the Series B round are underperforming Bitcoin since then, in addition to having tokens locked and vested.

On Tuesday night, Consensys announced that the SEC had notified them that it is closing its investigation into Ethereum 2.0. The SEC had alleged that sales of Ethereum constituted securities transactions. Consensys requested a letter clarifying Ether’s classification as a security, given the SEC’s recent approval of spot Ether ETFs, identifying the asset as a commodity. Although the SEC dropped its case against Consensys, Consensys is still pursuing its lawsuit against the SEC. Consensys noted that the investigation closure is a significant win for crypto, but they are seeking further clarity for the industry, attempting to get a court decision stating that applications built on Ethereum are not securities brokers, and therefore do not fall under the SEC’s purview. Crypto companies have had success against the SEC in court, and Consensys hopes to continue the winning streak and provide clear rules for decentralized applications.

Click HERE for more.

First News

All or Nothing. Here is a way to think about AI that might make someone’s life easier. Unless you are one of a tiny handful of businesses who know precisely what they’re going to use AI for, you do not need to have an active finger in the AI pot – or rather, you do not need to do anything to reap the benefits of such a stance. At this point in the hype cycle, artificial intelligence is likely already baked into the pie that is your businesses-software supply chain. Your managed security provider? They’re probably using certain algorithms, which they did not actively AI (to use it as a verb) either; they bought software from the tiny sector of the market that actually employs data scientists.

Strident pronouncements often use the phrase ‘make no mistake’, and so, to appear new and fresh, we will use a synonym. Let us endeavor not to err on this – generative AI does have the potential to drastically change the world. One scenario: we experience an explosion in silicon-based intelligence, where AI recursively self-improves itself until we are all living in the universe of the Matrix movie trilogy, in which a company’s bottom line will matter little, so we’ll pay a commensurate amount of attention to this.

A second possibility is a whimper-not-a-bang scenario. It may turn out that the current approach does not scale, as there’s a data bottle neck, the architecture doesn’t work, the algorithm stops getting smarter, context windows are limiting, etc. While some industries, such as customer support, will likely be heavily disrupted, your company will hardly need generative AI for the sake of it.

The final option is that we arrive at something that actually could replace programming and be identifiable as general intelligence. If generative AI goes on a hockey-puck – or rocket ship – trajectory (certainly what xAI is hoping for) then your business will be disrupted precisely as drastically as it would have been if you had done nothing (see last sentence of first paragraph, above).

^Scale.com

And so, to summarize: one way to look at AI now is that you either need to be on the bleeding edge of things, producing novel research, or you should be doing just what you were doing five years ago – with minor concessions to incorporating LLMs into your workflow. Anything in the middle makes less than average sense, unless you actually work in the rare field where your industry is being drastically disrupted and this very moment.

There’s gold in them there locks

There is another approach to AI, which gives businesses more agency. Although it was exciting for capturing our imagination, with the benefit of 20/20 hindsight, it is clear that AutoGPT was too general and, as a proof-of-concept for how LLMs might evolve to become general agents, it was too unconstrained to do useful things reliably.

Yet now there is a new cohort of agents out there that use custom cognitive architectures to provide guardrails that keep agents focused while also embracing the full power and capabilities of LLMs. There is a spectrum ranging from simple, hard-coded agents to full-blown autonomous agents. This framing shows us a happy middle where we are most likely to see useful agents emerge in the near to medium term.

Developers are finding their Goldilocks ‘just right’ balance in the middle of the agent spectrum, where they hand over a lot of the control flow to LLMs, but still maintain a set of rails and a sense of state.

The Goldilocks middle provides the best balance of power, flexibility and control, while being the hardest type of agent for developers to build, as it requires a structured but non-deterministic custom cognitive architecture. Simple agents are easy to code, since there is little stochastic variability to control for. Developing in the middle requires one to hand substantial control (and thus, stochastic variability) over to LLMs, while also controlling a high-level application flow and state management.

^Sequoia

Speaking of simple agents, there is a heuristic that Jeff Bezos is reputed to have given his employees, “Focus on what makes your beer taste better”. The analogy references how breweries at the turn of the 20th century made their own electricity, comparing that to how tech companies used to run their own infrastructure before AWS. In a world where agents fall over more often than not, where 12-13% on SWE-bench (a benchmark for evaluating LLMs on real-world software issues) is considered state-of-the-art, implementing a custom cognitive architecture just might help Pabst Blue Ribbon evolve into La Fin du Monde. Ludicity, Sequoia