I meme therefore I am
Joined June 2009
- Tweets 37,129
- Following 2,983
- Followers 9,215
- Likes 13,990
4,366 Photos and videos
Dan Sfera retweeted
$NVDA is trading near its cheapest valuation in a decade.
You’re paying a market multiple for the core compute engine of the entire AI economy powering the largest infrastructure buildout in history.
151
180
1,812
156,761
Still an excuse not to fix because not enough compute imo
I’ve had a number of conversations with folks inside and outside government about the current situation with Anthropic, and here is what I believe to be true:
— As we know, Anthropic publicly released its Mythos class models earlier this week under the commercial name Fable.
— Fable is Mythos with guardrails. But if those guardrails fail, then you’ve exposed Mythos and its advanced cyber capabilities to people who shouldn’t have them. (Keep in mind that Anthropic itself widely promoted the idea that Mythos was a cyberweapon and needed to be regulated as such. They asked for government regulation of Mythos and championed the guardrails on Fable. If there is a vulnerability — big or small — it is Anthropic’s responsibility to patch.)
— A highly credible trusted partner of both Anthropic and the USG who was testing Fable came forward with a jailbreak of those guardrails. The Admin asked Dario to fix the jailbreak or de-deploy the model. Dario refused.
— In their blog post, Anthropic defended its decision by saying the jailbreak isn’t serious. That is not what the trusted partner and the USG believe; nor is that kind of minimizing language consistent with Anthropic’s brand as the AI safety company. It’s difficult to fathom how they could claim a jailbreak allowing operability of a cyber weapon could be defined as not “serious.”
— In the past, Anthropic has always said that safety must be top priority and taken super seriously. In this case, Anthropic prioritized the continued offering of the consumer model over safety.
— In reaction, the Admin issued the export control. The Admin did this reluctantly. It’s been very surprised that Anthropic hasn’t wanted to cooperate with a reasonable safety request (ie fixing the jailbreak issue). Anthropic’s reaction is very much at odds with their branding and ethos as a safe AI research community.
— The Admin’s hope now is that Anthropic remediates the safety issue, the export control is lifted, and Fable goes back into general release. The Admin wants all of this to happen as soon as possible. It is frankly bewildered that Anthropic hasn’t wanted to comply with safety requests that it previously said were its highest priority.
— Those trying to misdirect and tie this action to the prior DoW/Anthropic issues are wrong. The Admin values Anthropic’s technical capabilities and feels that this issue, while serious, should be easily resolved. The ball is in Anthropic’s court.
252
Good Excuse to cover for lack of compute
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
1
308
Going Live, Playing Games Talking Research x.com/i/broadcasts/1OxwbbRAz…
194
Dan Sfera retweeted
Jun 11
I got a lot of flack when I said this before, but I continue to think that model providers will end up being a lot like airlines: Critical for the global economy but highly commoditized, with huge capex requirements and low (and often negative) margins.
Jun 11
OpenAI is considering drastic price cuts as it seeks to win over customers from archrival Anthropic on.wsj.com/4aldd0k
74
152
1,227
197,975
Dan Sfera retweeted
Jun 10
Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4)
182
571
6,022
3,445,517
spacex prices at a 1.8t tag tomorrow, 30 percent carved out for retail, book 3.5x oversubscribed. largest ipo ever and the float trades like a scarcity asset before it even opens. the part nobody is modeling is whether the xai merger turns a launch-cadence story into an ai one
1
202
.@pmarca: "The person who writes down the thing has tremendous power."
@david_perell
30
110
1,288
265,625
Dan Sfera retweeted
Jun 10
now that models like fable aren't subsidized anymore
its very bullish for companies building their own harness (devin, cursor, opencode, factory etc)
claude code becomes significantly less once useful since you're not "unlimited usage"
and these companies are highly incentivized to give you the best performance/ token (and have been trying to solve this for a while)
very good situation for companies all around (minus the users, cuz your ai bills are boutta 5x)
54
6
189
37,836
Dan Sfera retweeted
Jun 10
Most people can't actually explain what AI is to a friend/coworker.
Take a page from Dr. Seuss (trust me) & you'll be set.
When someone asks me "how does ChatGPT actually work?" I used to reach for words like "pre-training," "back propagation" and "transformers." And I'd watch their eyes glaze over (understandably) in real time.
So now I just say this:
If I say the words, "the cat in the ___," what comes next?
Your brain automatically fills in "hat," right? Mazel tov, you're an LLM.
An LLM is a machine that has read basically the entire internet, and all it does is predict the next word.
Over and over. "The cat in the" → "hat." Then it looks at "the cat in the hat," predicts the next word, and the next one, and the next one.
It's not thinking. It's not reasoning the way you do. It's the most powerful autocomplete ever built, so good at guessing the next word that the sentences it strings together can write your email, debug your code, or explain quantum physics.
That one reframe does two things for your team:
1) It kills the fear. "AI is coming for my job and I don't even understand it" turns into "oh… it's fancy autocomplete." Way less scary. Way more usable.
It builds the right instincts. Once people get that AI is just predicting the most likely next word, they suddenly understand why it makes things up (it's guessing), why specific prompts get better answers (you're shaping the guess), and why it sounds so confident even when it's wrong (it's always just predicting).
2) You don't need your colleagues to understand the math. You need them to understand "the cat in the ___." There's no downside to going down the rabbit hole & understanding AI at a mathematical/technical level, but it's not necessary if a professional simply wants to understand it well enough to work well enough with it.
Steal the analogy. It's the fastest way I know to take someone from "AI scares me" to "ok, I get it."
27
12
99
13,375