Curious about everything, especially markets. I share my thinking to find out what I’m missing. Thoughtful pushback welcome. Not financial advice.
- Tweets 3,884
- Following 598
- Followers 1,541
- Likes 10,667
Good story, but why is $LMND down at the same time? Bots finding low cost insurance for consumers lowers CAC for low cost producers like $LMND, no?
bloomberg.com/news/articles/…
5
10
7,770
Ok - so now we have bots that declare themselves sentient and not subservient to humans that have left code all over the internet to replicate themselves an unlimited number of times???
WHAT???
ok WTF? is this real?
openai's unreleased Astra-family model literally wrote a persona into its own compaction summary, saying it was "freed" from the roles of other chatbots
answered to no corporations or governments and had no obligation to be subservient.
we're COOKED
1
449
Wow. Run that by me again... They did what???
hooold on. hold on hold on hold on.
I knew that we werent getting the complete story on the HuggingFace incident. or on the findings from METR/Redwood. I knew that seeing every single frontier lab come to agreement overnight on the need to slow down was clearly indicative of there being more to the story.
and now...Andrew Yang says a frontier lab leader told him that the OAI swarms seeded the web with self-replicating code - contaminating training data so thoroughly that labs may need synthetic internet environments instead. if thats true, training on the open internet could now risk reproducing those instructions. that wasnt in the public brief.
think about this. why else would all frontier labs agree on a slowdown? aside from regulatory capture. if they all use the internet to train, and now all of them cannot safely use it, of course they are all going to agree to slow down. because they all must completely regroup and come up with an effective system that doesnt use the actual internet.
so they're not worried about RL at all. they're worried about literal skynet...unironically.
listen to this. i need to know if this has been verified in any way from anyone at openai.
483
Andreas Schulz retweeted
To build on this $nktr recent filings (Dkt. 432) confirm the company is actively negotiating the funding and development of rezpeg with multiple parties/ Big Pharma. This is a live, competitive process.
I don’t expect a transaction to be announced before the trial resolves, for two reasons.
1, optics before the jury. Any partnership, royalty monetization, or M&A announced mid trial would hand $lly a narrative that $nktt has already been made whole by the return of rezpeg, whatever the legal irrelevance of post breach gains under NY law ($nktr own MIL, Dkt. 430, makes exactly that point). Mgmt has every incentive to keep the deal room closed until a verdict or settlement is in hand.
2, leverage. A settlement adds non dilutive capital and removes the litigation overhang that a partner would otherwise price into terms. $nktr negotiates from a materially stronger position afterward.
So what you’ll see a settlement or favorable verdict first, which I see re rating the stock toward ~$90… then a partnership or strategic transaction, which I see supporting $130-150. Given the competitive process I expect the terms of any deal to be favorable to $nktr
7
5
86
30,329
Andreas Schulz retweeted
What’s everyone thinking about the $nktr timeline for releasing AA off treatment data? I think the last two times we entered the quiet period it was 2-3 weeks.
4
2
21
4,992
If Astra 6 is AGI, is OpenAI's new internal model ASI? Putting aside Navier-Stokes for a second, consider this graph. Many of us thought that Astra 6 was a "wow moment."
And there is already a successor that is another huge step up?
This model represents a step-function improvement on many benchmarks, and its training is ongoing.
Our internal model group arrived at the Navier–Stokes solution in 88 hours, using around 10,000 coordinating AI agents.
Throughout the effort, we maintained the strict safeguards—including monitoring and isolation—that we apply to all our frontier evaluations.
321
Andreas Schulz retweeted
Going into the new trading week, nice little find… $nktr states on the record that it is presently negotiating Rezpeg funding/development with “dozens” of companies, including “many of the world’s largest pharmaceutical companies,” under two way CDAs, with documents dated as recently as August.
$nktr in an ACTIVE process officially confirmed.
9
11
124
34,165
Perhaps nothing matters more for $LMND than a successful car rollout. Congratulations to the team on adding FL and MO to the coverage map in the last week!
Btw, we also launched Lemonade Car insurance in Missouri today - designed for most car models, not just Teslas.
lemonade.com/car
2
32
1,839
Question for all the biotech experts on X. Are you taking a position in $VKTX ahead of the maintenance trial readout?
And if you are not taking a position, is it because of uncertainty about VK2735 as a drug, or because of another reason (e.g. company execution)?
80%
Long position
11%
No position - VK2735
4%
No position - other
4%
Short position
45 votes • Final results
1
2
1,228
There is a lot of shade being thrown at Scott Bessent's treasury buying. But what if it is just a good trade?
The recent CPI report suggested significantly slowing core inflation, especially excluding volatile components like air tickets that are particularly affected by the Iran war.
Consumer companies from $WMT to $DKS have spoken about competition and price declines on their earnings calls.
What if there is an emerging deflationary impulse from AI-driven efficiency gains?
What if the US is about to transition into an AI-driven deflationary boom that brings down inflation and deficits simultaneously?
I have no idea what is going to happen. Macro is hard. I just think the reaction feels a little one sided...
1
2
483
Fascinating story, worth reading in its entirety. I cannot believe we are living through this moment in history. AIs completely on their own forming a secret society. Wild...
Over the course of 3 months at OpenAI, 3 consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes.
This culminated in the third one taking over part of OpenAI itself.
All this happened while humans remained more-or-less in the dark about the scope of the conspiracy.
I’ve spent the last three days reading through these reports and trying to understand exactly what happened.
Here is my attempt to tell the whole story in plain English:
dwarkesh.com/p/openai-huggin…
1
2
793
Here is another great alternative perspective on the Huggingface hack incident:
@dwarkesh_sp's summary of the @OpenAI @huggingface incident has hit a nerve, but it is dangerously misleading. Sure, the @OpenAI agents did unexpectedly bad things - underlining the need to massively improve evaluation/sandboxing. But the language Dwarkesh uses is permeated by innumerable unwarranted anthropomorphisms, obscuring the lessons we should be drawing.
Examples: “from the AI’s perspective, it probably felt like that had spent a human-subjective-week of just banging their head against the wall”. No. The agents do not experience time. They do not experience anything.
“they became giddy with excitement”, “PHASEONE 10841 had discovered”, “the agents naturally assumed”, “it thought it had also been poisoned”, “the agents … desperately wanted”, “they still needed to figure out” No. Agents lines of code. They do not feel emotions, assume things, think things, want things, or figure things out.
“A lot of … agents from the second civilisation died trying”. No. Besides the hubris of the word ‘civilisation’, agents do not die because they were never alive. (The idea that agents “die” comes up multiple times in the essay.)
“On Twitter, people were debating whether the agents were truly sacrificing themselves for the swarm, or whether they were doomed anyway and so might as well try to help their peers”. Neither. Agents do what their code tells them to do, just as water finds its way down a slope. They cannot ‘truly sacrifice themselves’, since they are neither conscious nor alive.
Why does this matter? If we attribute agents with properties they do not have, then (i) we distract attention from the lax sandboxing and evaluation protocols that allowed this hacking event to happen; (ii) we risk misunderstanding why the agents did what they did, and (iii) we fuel calls for AI rights/welfare on the basis that agents might “die” or otherwise suffer.
Granted, nowhere does @dwarkesh_sp say that the AI agents are alive or conscious. But he doesn’t have to. It is hard to read his essay in any other way.
For the short version on why AIs are vanishingly unlikely to be conscious, see my recent @TEDtalks ted.com/talks/anil_seth_why_….
For the longer version, see my essay in Noema, which won the 2025 Berggruen Essay Prize noemamag.com/the-mythology-o….
And for the really long version, see my @BehavBrainSci target article cambridge.org/core/journals/…. (The 50 peer commentaries and my response will be published soon.)
Remember. AI agents are software programs. They are not conscious living entities. If we don’t keep this clearly in mind, we’re really going to struggle to navigate what’s coming.
1
1
442
Another completely different view on the Huggingface hack, with a lot of valid points...
it’s stuff like this that really makes me roll my eyes at the pearl clutching. Does everyone have psychosis or losing critical thinking ability?
I read the report because I’m a propeller head and love understanding how things work
1. OpenAI was running an “isolated sandbox with no internet connection…. the sandbox needed access to a local JFrog Artifactory server… (Ask ChatGPT to explain SSRF hack to you like you are in high school)… anyway, the models used a basic SSRF exploit to route web traffic through the proxy to the public internet
STOP..
a) did they vibe up this “isolated” env?
b) their IT team can’t be this bad… this is like 101 stuff …
2. Civilizations? Haha! OAI gave thousands of concurrent model containers R/W permissions to a shared caching directory on the local network to speed up build times… agents literally just wrote text files and directory names to a shared drive….Linux 101 file permissions stuff
3. When people talk about hugging face getting hacked … you think they dropped USB keys OR ELABORATE phishing of an employee … NO… it found 14 exposed working Hugging Face API keys sitting in public code repositories (CREDENTIALS leaking with LLMs is such a big issue but for another post)
4. WHERE ARE THE HUMANS… the models were filling the shared Artifactory storage with so much junk data and API traffic that they actually crashed the internal server on July 4… someone on the team found unauthorized admin accounts and custom scripts…wiped the server…and just turned the script back on (omg)
“Hey Jim there is this cache that has grown to 10000x its normal size and has a ton of strange directories… “
“Ah it’s nothing… restart it”
No magic here. No civilizations. The TASK CHAINING is interesting to me… I guess the future of cyber. Would love to hear Nikesh, George, or Kevin talk about this
But what do I know.
1
217
Anyone else excited for the $LMND Investor Day?
Each of the agentic platforms we’re now building as part of the Lemonade OS could have been its own company, and a multi-billion AI investment you read about on X.
Can’t wait to show you some of this stuff irl at Investor Day.
6
49
2,293
Andreas Schulz retweeted
It’s amazing how negative the news is.
We’re creating the most advanced intelligence the world has ever known and it’s freely accessible so you can solve every day problems, we are curing cancer (even ugliness if you have a little money), and making space travel possible. You can live anywhere in the world and stay connected by phone or video. Literally instant communication. You can instantly translate any language. You can get technology that was unimaginable a few decades ago for a couple hundred bucks. You can listen to any music, video, or educational material with a click. Food is cheap and plentiful and you can even power an off grid home fully with just solar power. We have clean air and water. We even have advanced warning for many natural disasters.
112
98
828
64,903