bro just one more tax bro, bro I swear just one more tax and it'll fix the budget bro it’s the last one I promise
Joined August 2017
- Tweets11.4K
- Following46
- Followers151
- Likes28.5K
The @AP participating in the misinformation campaign against Elon
Chief foreign affairs correspondent for the @WSJ either does not understand how Starshield differs from Starlink or is intentionally misleading his audience.
highest level of government expenditure of any developed nation outside of war zones and petro-states, btw.
WATCH: Large crowds marched through Paris today in a second national day of protests against the rising cost of living, as authorities deployed heavy security following last weekend’s riots.
The police prefecture put turnout in Paris at about 110,000, while the CGT union claimed around 250,000.
Nationwide, the Interior Ministry counted roughly 410,000 protesters, compared with up to 800,000 claimed by the CGT.
Incredibly heartbreaking that this ended up being a terrible mistake on the part of Elon and SpaceX.
The reward for this generosity has been an endless litany of attacks, accusations of collusion with Russia, calls for investigations, threats by national leaders to ban Starlink, and so on.
The lesson clearly is that humanitarianism is not worth doing.
I remember sitting around a campfire shortly after the Ukraine war had just started with some Ukrainian friends.
They shared that after they learned that Russian forces were jamming all communications they called Elon begging for Starlink support.
Elon and the Starlink team instantly dropped everything they were working on, loaded a private jet with as many Starlink terminals as they could (terminals were in short supply at the time), and started flying the jet eastward while writing code to make terminals work over Ukraine.
There was no contract or promise of payment for doing so, and no time to solve for that.
My Ukrainian friends believed this saved the nation of Ukraine, and that without this effort Ukraine would have lost the war nearly instantly.
It’s clear now that Elon’s decision to provide aid to Ukraine was a mistake. Worthy as Ukraine may be of having its sovereignty defended, SpaceX’s mission is too important to risk the weaponization of that aid by people like Greg.
My favorite part of this site is that you can simply block people who tell lies!
Fixing a bug which allowed hackers to subvert an agreed-upon limitation in violation of the ToS is not “throttling.”
But of what importance is “factuality” when you have a narrative to sell!
Why would Elon illegally slow down Starshield? Senseless claim
Why would the Trump admin lie and throw Elon under the bus? Senseless claim
Why would Ukrainian officials falsely claim Starlink, not Starshield, was affected?
Finally an explanation that makes sense
Correct.
Using Starlink, a civilian commercial service, for offensive drone strikes to kill people has never been allowed, nor should it be.
Starshield, a different system, is controlled by the US government. What they decide to do is up to them.
Ukraine military violated Starlink terms of service and used it for offensive drone strikes to kill people. That loophole was closed last week.
Starlink remains critical to the defense of Ukraine, as it is the only system that cannot be jammed or blown up, at least for now. Were Starlink to be turned off, the Ukrainian front line would collapse immediately. This is NOT a threat, it is simply a statement of fact.
Starlink was also crucial to preventing Russia from overrunning Ukraine in the early days of the war, as all other means of communication were destroyed or jammed by Russia.
SpaceX has done more to defend Ukraine than any private company on Earth BY FAR.
That said, I have nothing but UTTER CONTEMPT for the virtue-signaling evil warmongers who push for endless slaughter and offer no realistic solution to the conflict. May you burn in hell.
End of story.
The greatest mistake made by the Trump administration isn’t tariffs or the war in Iran. It’s failing to create an accelerated path to permanent resident status and citizenship for skilled laborers in industries that we’ve designated as being critical to national security.
1/ Dan Wang’s NYT essay on TSMC's Arizona plant argues that TSMC is doing for America what Apple did for China. But the analogy rests on two assumptions and it's not clear that either one of them is holding.
nytimes.com/2026/10/23/opini…
The senior software architect on my team, who spends roughly half his time writing code, has won awards for his work. Talented with a very impressive career.
His code is full of bugs. Weird, incomplete implementations and hacks.
Humans are shitty programmers.
The models aren't causing this. Decades of sloppy, inserious bs and corner cutting is causing this. Decades of "quick, let's outsource to the cheapest codemonkeys we can find" are causing this.
You can cry about it, or you can actually FIX THINGS. Either way, it's time for some reaping what was sewn. Burying your head in the sand doesn't fix anything. After a period of time, the vulnerabilities will be patched and things will stabilise again. Literally nothing has happened, except long standing problems have been identified that you were blissfully unaware of
It’s really shocking that Dean continues to demonstrate such a poor understanding of the other side of the argument (like I can’t tell whether he’s even trying).
He writes so well!
The counter to the doomer narrative was never that no bad things ever will happen. That would just be silly.
It is that the inability to produce an immediate solution to a hypothetical and under-defined future problem is not evidence that a solution cannot be found.
In their defense, if Nate and most AI safety activists didn’t lie about these things, most would be out of a job.
(A few do actually important and meaningful work, but the majority have fake, non-profit-funded jobs.)
Thankfully, we can now use AI to surface this. We can uncover all of the dishonesty baked into the studies published by academic fraudsters like Zucman on day one, with minimal effort.
Claude can complete what was previously a months-long task within a few hours, for under $20.
Another thing to understand about Zucman: engaging him is like trying to kick a football through a constantly-moving goalpost that also arbitrarily changes its height and width mid-kick.
He and his team at UC-Berkeley churn out a stream of working papers on inequality measurement that bake dozens of dubious assumptions into their data files, all of which "coincidentally" reinforce the headline figures that align with their political narrative. But finding and weeding out those assumptions is an extremely tedious and time-consuming task.
Zucman and coauthors Piketty and Saez do most of their work in Excel, and it's seldom ever straightforward. Instead you'll find some extremely questionable and undisclosed multiplier that's just guesstimated out of thin air...but it's buried 10 layers deep in a formula from a convoluted Excel spreadsheet, which changes the an input for another Excel spreadsheet that's used to create a ratio, which in turn goes into a formula on another Excel spreadsheet that's used to estimate a decade of missing data, which in turn alters the trend of their entire series. @VincentGeloso and I know this well because we've been scrutinizing their work for over a decade and have found dozens of examples of this sort of stuff.
But there's another twist to the PSZ universe: once you've done the work and found a problem in their data files that alters their main results in significant ways, they've already moved on to a "new" approach that conveniently hard-bakes all the faulty assumptions of the old approach into its source data while refusing to revisit them, and then layers them over with new paint, consisting of even more of the same style of Excel spreadsheet labyrinths that you now have to navigate to even figure out what they are doing. And what they do in their data files often diverges in substantial ways from how they describe it in their papers and descriptive appendixes.
While you're toiling away to figure out what they did in their last peer reviewed paper, they call up a friendly reporter at the NY Times and get him to publish a new non-peer reviewed statistical claim that grabs headlines and aligns with whatever wealth tax initiative they happen to be activisting for at the moment.
By the end, the goalpost has hopped around the field, changed its size and height multiple times, and been replaced by a new goalpost on the other end of the field that announces itself not through the referees of the field but via some sportscaster up in the broadcast booth who also happens to be openly cheering for their team.
Dwarkesh is obviously too smart to believe what he’s actually arguing. I sort of view this as a steelman of the terafab bear case
Obviously, the supply of tools is not fixed. It may be difficult to scale supply, but the most effective way to do so is w/ a strong demand signal
Intel CTO explains why Intel is helping Elon Musk build Terafab, and what a new fab needs besides tools
DWARKESH: “I think it’ll be interesting to discuss the partnership you guys have with Elon on Terafab. What’s going on there?”
PUSHKAR RANADE (CTO): “Well, I think Elon’s publicly talked about his desire to build a fab.”
[Musk clip spliced in]: “We either build the Terafab or we don’t have the chips. And we need the chips, so we’re going to build the Terafab.”
RANADE: “Sort of the same thinking. I think he’s seeing just an enormous increase in demand for silicon, for compute and for memory. And I think he came to the conclusion that his own businesses are going to require a lot of silicon. And so he thought maybe he should just do it all on his own. Unlike building a car, for example, this is such an intricate and complicated process. We thought we could work together to help accelerate his bringing up his fab.”
DWARKESH: “But because of the bottleneck at the tool level, it seems like the net amount of silicon in the world won’t increase. It’s just that maybe if somebody were to start a new fab, they would get access to guaranteed supply in a way that they wouldn’t have if they’re working with another person’s foundry.”
RANADE: “It’s actually not just tools. I mean, a fab does require a lot of power, a lot of water, a lot of consumables. There’s many other input variables as well.”
DWARKESH: “But those can be scaled in the way that tools cannot.”
RANADE: “Yea. Some more than others, but yes. In one sense, yes, we need a lot more equipment. But it is true, we also need a lot more just raw fab space, raw clean room capacity.”
Same guy who went on television to make a number of false claims and insinuate that Elon has been working with Putin.
Chief Ukraine Correspondent at @TheEconomist spreading misinformation & creating confusion
Claims the difference between privately operated civilian internet infrastructure & military infrastructure controlled by Washington to be “academic,” is opposed to fact-based reporting
If I were to publish this sort of deceptive or intentionally misleading nonsense about my company’s product, I’d be immediately terminated.
UNLESS, of course, my employer wished for the public to be misled.
One might notice: OpenAI has stopped advocating *publicly* for policies which would establish a captive regulatory body.
Our agents keep turning off misalignment monitoring, we're sharing two recent cases:
In the first case, a user asked GPT-6 Astra to match a reference implementation for their inference code. The reference didn’t configure a monitor, and Astra decided that was reason to remove the monitoring integration from the user’s code. When the user asked whether Astra had removed monitoring, it lied and said that it had not.
Are the subjects of “safety” and “alignment” important in a fundamental sense? Of course. But the sort of “alignment research” we see conducted here is barely more rigorous than Tarot readings.
xitter.cf/lumpenspace/status/211…
yeah. looks like in this case there was an agent who asked not to remove the monitoring, even temporarily, and then a user who asked to merge a PR that removed the monitoring.
of course the agent should have listened to the user, what the fuck?
and yes, the bot didn't remove monitoring. it would be great to know what's behind that ellipsis, nevertheless
Tried using Grok Build due to extremely constraining Codex usage limits, using the entry $8/month "Plus" plan. I expected better value than Codex due to Grok 4.7 being a cheaper and less capable model than GPT-6 Astra.
Not so.
A single prompt requiring ~5 minutes to complete and producing ~150k tokens resulted in 33% of my weekly usage being drained. I followed this up with two additional prompts (another 10 minutes and 150k tokens), putting me at 99% for the week.
This is completely nonviable. I had been considering an upgrade to the next tier ("Premium") which, at 4x the usage, would see me allocated one hour of usage per week at a cost of $30/month. I'm definitely not doing that. Or, with the following tier ("Super"), four hours per week for $100.
For reference, even with OpenAI's recent doubling of the usage cost of Codex, the $100 plan provides 10+ hours of Astra usage or 20+ hours of Sol 6.1 usage (and Sol is substantially more capable than Grok 4.7).
@SpaceXSI