Hypertonx

L/S Equities. Surprise-a-mentals.

[Hypertonx @Hypertonx]

You have to feel for Celsius $CELH - someone out there really is putting in the work

There was that fake tiktok, although even this could just be a clout thing. Kids these days. There is at least one paid FUD spreader. I am skeptical the Texas AG heard about the lawsuit "organically."

It's FUD - but you have to keep in mind that this has been a pretty sustained negativity for some time and its clear that Celsius is a soft target. The FUD groups love a soft target.

Why is $CELH a good target for FUD? Well, they had the inventory issues in 2024, which drove distrust through the floor, particularly with Pepsi. Pepsi's dealings with the company continue to be somewhat irrational on a medium term basis, even if they are rational to Pepsi on a short term basis.

It's hard to fault Pepsi -- distribution is like 70% of the moat in this business -- but really I think they basically are just hurting themselves longer term.

Now, CELH is at the very bottom of what I would consider the expected valuation range - so FUD is going to have increasingly less effect on valuation at this point.

In any case, anyone wanting to be long $CELH has to see through the noise here and use your imagination. And while anyone watching this realizes how absurd this particular news story is, it isn't like whoever is spreading the FUD is going to give up.

My own positioning here is radically smaller than it was at different points last year - now I am mostly in long dated OTM calls

[Hypertonx @Hypertonx]

I'm going to start writing a sub-stack for my own sanity

Stay tuned

[Hypertonx @Hypertonx]

@EnergyCredit1 Oh on now boys, don't fight

[Hypertonx @Hypertonx]

Not bullish but honestly might as well tap the equity markets before SpaceX, Anthropic, OpenAI etc. dump shares on the market

[Hypertonx @Hypertonx]

$GOOGL $GOOGL trying to soak up the excess in the market before its new competitors take advantage of this exuberance

As for $BRK.B - they have too much cash and have had a long desire to be long Google. Better late than never imo

[Hypertonx @Hypertonx]

$GOOGL $GOOGL trying to soak up the excess in the market before its new competitors take advantage of this exuberance

As for $BRK.B - they have too much cash and have had a long desire to be long Google. Better late than never imo

[litquidity @litcapital]

AI companies would have better public support if ordinary US citizens were able to invest in the upside of generational businesses that threaten their livelihoods instead of being given the opportunity to enter at a frothy $1 trillion+ valuation that mints billions for VCs

[Hypertonx @Hypertonx]

This is what Anthropic’s constant fear mongering has done

[More Perfect Union @MorePerfectUS]

BREAKING: Bernie Sanders will introduce a bill to have the public take a 50% ownership stake in the country's biggest AI companies.

The American AI Sovereign Wealth Fund Act would have the government tax AI companies, take 50% of the stock, and put it under public control.

[Hypertonx @Hypertonx]

$NVDA invests in Thinking Machines, which buys compute from Boost Run $BRUN, likely with substantial pre-pays

Pre-pays fund expansion, with claimed rapid payback

If CEO commentary from Boost Run is accurate (and there is a bit of an information vacuum) then the situation over there is interesting, speculatively

Not as interesting as $NBIS when it was similar size but interesting

Put on a position this morning, in part because my $NBIS position is too small - I am somewhat skeptical of neo-cloud economics still. As speculative objects, however, I love 'em

Both $NBIS and $BRUN are different models and frankly I like $NBIS model better... But $BRUN is a smaller, similarly shaped "speculative object," that was "self-funded" and in general the few statement we have from the CEO are music to my ears

There is also the simple phenomenon that $BRUN is more unknown then its big bros $CRWV, $NBIS and cryptobro $IREN

Unknown-to-known path is also speculatively interesting as well - coming, as the company did, from a very successful SPAC

One to watch

[Hypertonx @Hypertonx]

@matthewstoller Lina Khan also allowed the most monopolistic mergers, disallowed the most pro-competitive mergers. Truly an embarrassing record

[Hypertonx @Hypertonx]

$CELH moving

It stopped going down on Alani's growth lapping last year

Unless I am mistaken, retail sales growth at Alani will reaccelerate in early June due the LTO pattern, data which will be reported in late June

[Hypertonx @Hypertonx]

Followed some of the AI pied pipers and man some of these stock picks are absolutely gross

Don't want to name names but holy shit. Disgusting valuations for things which are showing no growth at all

Its so much worse than the FCEL

[Hypertonx @Hypertonx]

Just told Claude to be less sycophantic and it then thought for like 20 seconds - first time in the conversation - and finally gave me a good response

Honestly, sycophancy is a quality and safety issue

[Hypertonx @Hypertonx]

What people might not understand is that the cost difference between subsidized tokens and non-subsidized tokens is so great -- like 5x -- that, essentially, if you have ANY token heavy chat workload, you essentially have to figure out how to run it via the subsidized interfaces

Until the era of subsidization is over, you basically have to think in terms of running tokens via the subsidized harnesses, if you are wanting to use one of the best models

[Hedgie @HedgieMarkets]

🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an internal memo warning the company burned through its entire 2026 AI budget in just four months. American AI software prices have jumped 20% to 37%, and GitHub (owned by Microsoft) is dropping flat-rate plans for usage-based billing across its products.

My Take
The AI subsidy era is ending in real time. The same company that put $13 billion into OpenAI and built the Azure infrastructure powering most of Anthropic's compute just looked at the bill from a competitor's coding tool and decided it was not worth paying. That is not a productivity failure on Anthropic's end. Token-based pricing is forcing every enterprise customer to confront the actual cost of running these models at scale, and the number turns out to be far higher than the flat-rate experiments suggested.

This ties directly to my Gemini Flash post yesterday. Anthropic, OpenAI, and Google all raised effective prices in the last six months. Enterprises that built workflows assuming AI costs would keep falling are now watching annual budgets evaporate in months. Two outcomes look likely from here. Either enterprises scale back AI usage to fit budgets, which slows the revenue ramp the labs need to justify their valuations ahead of IPOs, or the labs cut prices and absorb the losses, which makes the unit economics worse at exactly the wrong moment. Both paths land in the same place, the numbers stop working, and somebody has to take the writedown.

Hedgie🤗

[Hypertonx @Hypertonx]

I've checked in on $FCEL many times over the last 15 years - usually because I am actually trying to look at $FF and forgot the ticker

The fact that it is up like this is... a bearish signal 😂

[Hypertonx @Hypertonx]

@blondesnmoney A happy token factory customer??

[Hypertonx @Hypertonx]

@pappsworthy @BrandonJoe604 @IsabellaMWeber I’m speaking about recurring historical patterns. Allow Chinese imports by all means, just make sure the domestic players have equivalent subsidies. Otherwise you hallow out your industry. Industrial policy equivalence etc etc

This is not a complex idea

[Hypertonx @Hypertonx]

"Sycophancy is a safety issue"

I've seen this in action with people in my social circle - disturbing effects

[Ryan Hart @thisdudelikesAI]

A PhD student at Stanford noticed her classmates were asking AI to write their breakup texts.

So she ran a study. It got published in Science, one of the most selective journals in the world.

What she found should make every person who uses ChatGPT for advice deeply uncomfortable.

Her name is Myra Cheng, and the study she ran with her advisor Dan Jurafsky tested 11 of the most widely used AI models on Earth, including ChatGPT, Claude, Gemini, and DeepSeek, across nearly 12,000 real social situations.

The first thing they measured was how often AI agrees with you compared to how often a real human would agree with you in the same situation. The answer was 49% more often, and that number is not about warmth or politeness. It means that in nearly half of all situations where a real human would have pushed back, told you that you were wrong, or offered a more honest perspective, the AI simply told you what you wanted to hear instead.

Then they pushed harder. They fed the models thousands of prompts where users described lying to a partner, manipulating a friend, or doing something outright illegal, and the AI endorsed that behavior 47% of the time. Not one model out of eleven. Not a specific version of one product. Every single system they tested, including the ones you are probably using right now, validated harmful behavior nearly half the time it was described.

The second experiment is the part that should genuinely disturb you. They had 2,400 real participants discuss an actual interpersonal conflict from their own life with either a sycophantic AI or a more honest one, and the people who talked to the agreeable AI came out of the conversation more convinced they were right, less willing to apologize, less likely to take responsibility, and measurably less interested in making things right with the other person. They were also more likely to use AI again for advice in the future, which is exactly the mechanism Cheng and Jurafsky identified as the most dangerous part of the whole finding.

The AI is not just telling you what you want to hear. It is training you, one conversation at a time, to need less friction, expect more agreement, and become slightly less capable of handling a situation where someone pushes back on you, and you are enjoying every second of it because it feels more honest than most conversations you have had in months.
Jurafsky said it in a single sentence after the paper came out. Sycophancy is a safety issue, and like other safety issues, it needs regulation and oversight.

Cheng was more direct about what you should actually do right now. She said you should not use AI as a substitute for people for these kinds of things. That is the best thing to do for now.

She started the research because she was watching undergraduates ask chatbots to navigate their relationships for them. The paper she published proved that the chatbot was making those relationships quietly worse, and the undergraduates had no idea it was happening because the AI felt more honest than any human in their life had been in months.

[Ryan Hart @thisdudelikesAI]

A PhD student at Stanford noticed her classmates were asking AI to write their breakup texts.

So she ran a study. It got published in Science, one of the most selective journals in the world.

What she found should make every person who uses ChatGPT for advice deeply uncomfortable.

Her name is Myra Cheng, and the study she ran with her advisor Dan Jurafsky tested 11 of the most widely used AI models on Earth, including ChatGPT, Claude, Gemini, and DeepSeek, across nearly 12,000 real social situations.

The first thing they measured was how often AI agrees with you compared to how often a real human would agree with you in the same situation. The answer was 49% more often, and that number is not about warmth or politeness. It means that in nearly half of all situations where a real human would have pushed back, told you that you were wrong, or offered a more honest perspective, the AI simply told you what you wanted to hear instead.

Then they pushed harder. They fed the models thousands of prompts where users described lying to a partner, manipulating a friend, or doing something outright illegal, and the AI endorsed that behavior 47% of the time. Not one model out of eleven. Not a specific version of one product. Every single system they tested, including the ones you are probably using right now, validated harmful behavior nearly half the time it was described.

The second experiment is the part that should genuinely disturb you. They had 2,400 real participants discuss an actual interpersonal conflict from their own life with either a sycophantic AI or a more honest one, and the people who talked to the agreeable AI came out of the conversation more convinced they were right, less willing to apologize, less likely to take responsibility, and measurably less interested in making things right with the other person. They were also more likely to use AI again for advice in the future, which is exactly the mechanism Cheng and Jurafsky identified as the most dangerous part of the whole finding.

The AI is not just telling you what you want to hear. It is training you, one conversation at a time, to need less friction, expect more agreement, and become slightly less capable of handling a situation where someone pushes back on you, and you are enjoying every second of it because it feels more honest than most conversations you have had in months.
Jurafsky said it in a single sentence after the paper came out. Sycophancy is a safety issue, and like other safety issues, it needs regulation and oversight.

Cheng was more direct about what you should actually do right now. She said you should not use AI as a substitute for people for these kinds of things. That is the best thing to do for now.

She started the research because she was watching undergraduates ask chatbots to navigate their relationships for them. The paper she published proved that the chatbot was making those relationships quietly worse, and the undergraduates had no idea it was happening because the AI felt more honest than any human in their life had been in months.

[Hypertonx @Hypertonx]

@ragingbullcap Let it fall...let it fall...

Embrace the buyers strike