Chris Hawkins

@tracewoodgrains @MichaelKiwi77 @it_sted Blue origin is very good for the world if you believe a competitive space economy is important.

yeah, I don’t think Amazon’s success meaningfully depended on Scott, or that there was ever particular reason to expect her to effectively steward a large chunk of Amazon. So her substantially reducing her power over it is basically fine but she’s giving the power to many I suspect will use it poorly. Bezos is mostly keeping his around or building rockets, which — cool, not the best thing that could be done, not the worst

You can tell quite a bit about someone by whether they think MacKenzie Scott or Jeff Bezos has been a more responsible steward of their billions.

@techniahqrobot Robot “customer” too. Not even dancing for humans 😂

Humanoid robotics has officially entered its pole dancing phase. This prototype performs a choreographed full body sequence while gripping a fixed support pole and moving on a rotating platform. The hip knee and torso coordination is interesting. But the rig also means this is not proof of unsupported balance or autonomous locomotion. Serious motion control experiment or attention engineering?

@tunguz Who drinks regular coke? Diet Coke above all.

I'll never get over the fact that the best Coca Cola - THE most iconic American product - is made in Mexico.

I suspect the issue is that by definition these models are supposed to be generally intelligent, meaning they can be swapped easily. So what does “product differentiation” even mean? Most usecases do not need more than Opus level intelligence if they have the right scaffold. This includes most tasks in legal, accounting, property management, etc. BUT to actually replace lawyers accountants property managers etc there is a bunch of plumbing stuff that will need to be done by service companies or app layer companies. And the LLMs they need, (Opus 4.6 level intelligence) are already getting commodotized, meaning only infra will make money from most usage soon. This means the labs will have to focus more and more on cases with nearly infinite demand for intelligence, like drug discovery and development.

If you abstract away the fact that this is AI, and you solely looked at token price and volume as an output of rising infrastructure costs, wouldn’t the only logical outcome eventually be that one of the models figures out a way to commoditize the core complement(s) (currently memory => inference) to become the low cost provider while the remaining models fight it out on product differentiation?

The mega bull case for AI infrastructure would be *if* market share shifted away from certain frontier labs with 90%+ inference margins toward cheaper models, whether open-source or closed. It would increase the ROI on AI spend for end customers by increasing intelligence per dollar, which would drive incremental token demand. Margin dollars would effectively get redistributed from the frontier labs to AI infrastructure providers. The infra winners would be those with the lowest per token cost and the winners at the model layer would be those with the highest token efficiency. There are many reasons Jensen is so focused on open source, but this is likely the most important one as I think he is probably less worried about a monopsony these days. Lower margin % at the model layer = more margin $ at the infra layer all else equal. With SpaceX and Meta being vertically integrated and possessing the #3 and #4 models respectively it is more possible than ever. Note that Grok 4.5 is ahead of Fable for some useful tasks at a much lower cost, so ranking them #3 is conservative. This is not happening yet. Cheap, mostly open source tokens are likely the majority of volume today but the majority of economic value is still accruing to the most intelligent models. Might change though. We will see.

Self driving cars/taxis can do point to point drop off without parking, making them will be vastly superior to everything other than planes at distances up to 200 miles in pretty much every case, and further in many more. If we really wanted to support public transit we would try to increase competition and reduce the cost of transit providers.

@mcuban Love costplus. Genuine question: who does a good job here? Like do you have a model company/setup you can point to as a guide?

It's not. Why ? Because 25 pct or more of a doctor's time is spent dealing with conglomerates that do all they can to make the doctor's care more difficult, and expensive, for both the doctor and patient. For every future agent we give AI doctors to deal with this friction, and to improve the quality of care, the conglomerates will have multiple adversarial agents doing all they can to delay and deny, to minimize their cost and maximize their float We see this already as the conglomerates use AI to find every possible way to manipulate contracts, and find ways to mislead, while hospitals hire companies for Revenue Cycle Management, who charge as much as 10 pct of revenue to have their agents try to do the reverse. It's the agentic version of Mad magazine Spy vs Spy I'll give you a further example. There isn't a single company, including yours, that knows the actual cost of the care they purchase for your employees and families. Not one. Cost is an important component of health care decision making. @a16z includes costs in defining its benefits. But you are blind to all but the total bill you pay. Your carrier, your ASO, your PBM, any company that touches the economics of care for your company is going to do everything they can to prevent you from using AI doctors or agents successfully If you want to see that change, stop working with the healthcare conglomerates. Write agents that define , optimize and contract directly with providers, to eliminate the uncessary middlemen. Feel free to use costpluswellness.com to train them. Until the conglomerates are disintermediated , HC in this country will continue to be fucked

@dhh There is no supercycle. We are in a forever war between those who seek to expand the pie vs redistribute the pie.

"The circle of the saeculum is both a prophecy and a roadmap. We're not supposed to live like this forever: weak, ineffectual. This too shall pass. We just have to make it out of the current Crisis alive." world.hey.com/dhh/the-will-t…

@allTheYud @AnthropicAI They keep resetting limits and expanding usage, so it can’t be that they are really compute constrained at this point.

@anthropicai, do not make the mistake the OpenAI board made. Explain your reasons fast. If the reasons are, "We are using training compute to serve Fable broadly and we can't keep that up indefinitely", SAY SO to avoid losing customer trust and respect.

@KenoFischer Safeguards a bit rough for anything close to security related, which I get but is irritating

But seriously, Fable is a fantastic model, absolutely love it, but dripping it out a week at a time is just such a weird thing to do.

GDP will go up over time. The cost of achieving 2026 level abundance will fall. People will raise the bar, continuing to demand new things, until demands are somethjng like middle class person complaining their island sized private space station is not as big as the richer persons continent sized space station.

thought experiment: every person on earth suddenly gets 7 AGI robots that are as strong as a strong human and as smart as a smart human + run on solar energy for free how much does global GDP increase?

@cremieuxrecueil I like Anthropics products and people, but their push of the Doomer narrative is awful. How about an ad “Social security will run out in 2032…AI can save it.” That is a good ad!

So is Anthropic just designing their ads to scare old people? Is this the plan? That ad with the person going 'What if AI can care for people better than I can?' was unhinged.

@scottlincicome Let’s actually pass this thing this time. Changing the clocks is horrible, just pick a time, any time, and leave it there.

🚨 BREAKING🚨 "Hate daylight saving time? New bipartisan House bill would repeal it" 🔥 news-journalonline.com/story/news/202…

I don’t think AI will cause mass unemployment but less than 50M people in the US are in atoms (rest in bits). Those in bits will see like 100x more competition from digital labor and the US will plausibly get a disproportionate share of early robots, meaning if they got 25% of 4B robots this would be a 20x increase in labor, not 2x. I think abundance is great!

Dr. Mike Israetel on the economic fallacy he says explains why AI won't cause mass unemployment: "Once we have 4 billion robots doing labor in the world, which we're like orders of magnitude off of that currently, then we've just only doubled the human workforce." "From 1700 to today, we've 10 or 20x'd the human labor force. And, seemingly, the economy's not like, ah, we don't need any more people, that's enough. We could just consistently have better jobs and pay people even more money." "This idea that robots are gonna show up and all of a sudden we're all completely unemployed makes a technical fallacy in economics called lump of labor fallacy. It's the idea that all the jobs currently are the only jobs that could be." "Imagine in 1750, you're like, well, 98% of us work in farming, and then you come back from the future and you're like, you guys, 2% of people in the 1990s work in farming. It'd be like, so everyone's starving to death? Like, no, no, we're super fat, actually." @misraetel

@nfthing_ @ZackPolanski Genuine question: what are you referring to? Property management companies? These costs are coming down fast (eg tidy.com charges 80-90% less)

@ZackPolanski Still no mention of the service charges by 'management companies' that keep rising & force landlords to increase rents - this is the key issue and is never mentioned.

Here's what the landlord lobby don't want you to know: If rents went down by 20%, renters would save £2400 a year. The government would save £2 billion a year. And landlords would still make 4.5 times as much profit as the average business.

Native Americans, being wiser and less wasteful than the colonizers, use every part of the space. Being more in touch with the land, they understand that you can actually build things on it.

@ramez They also simply ignore the benefits of AI entirely when considering delaying AGI from 2030 to 2040. About 700 million people will die, social security will run out, and the economy might be 10-50% smaller. AI is the current Plan A to fix these with no plan B.

I'm firmly in the third camp. The basic problem with AI 2040 is that it uses a fictional and speculative doomsday scenario to justify very real surveillance and control capabilities that governments would be certain to use in authoritarian ways, well beyond AI safety. It warns against concentration of AI power but its policy proposals serve to increase the power of the most powerful entities on planet Earth. It concretely sacrifices freedom in ways guaranteed to cause harm, in an effort to forestall a made up threat. It proposes safety tools that give governments unprecedented capabilities to monitor, suppress, and manipulate. These tools are intended only to stop the development of overly powerful AI, but once they exist, governments will use them as they please. Its authors mean well, but are so convinced of a fictional and unproven threat that they'd do real harm to the world to prevent it. It would create a world that is less free and less safe in the name of safety for a threat that may not even exist. I can't imagine making this trade off.

@zephyr_z9 There is also GPT-5.6 Luna. Same ballpark price benchmarks. Price war around opus level is on.

While I haven't seen muse spark performance on independent/private benchmarks But if the benchmark results are decent, then things are about get bad for Opus & below tier models Rumors of OAI starting a price war was circulating but Zucc and Elon triggered it

We should just not have pit bulls. Im not sure if it’s woke that keeps them around but we should not have them. I tell my child to avoid pit bulls when he’s out, categorically.

@jmbollenbacher @teortaxesTex The US is heavily services oriented, so 80% of the economy is bits not atoms. But part of AI push is in atoms. Self driving cars plus bipedal general robots end up removing lots of constraints. I think an incremental 1-5% GDP annual growth is defensible.

@teortaxesTex "bits are easier than atoms" is also why AI will not produce a 50% gdp growth rate or whatever nonsense people are peddling this week. itll stay more or less on trend. reallocation and redistribution, sure, but growth will still be capped by the speed of logistics.

Incredible that we'll have AGI before fusion

@ericzakariasson @maxedapps If you make a Deepseek v4 Flash killer, would be great as subagent and very useful for dev applications. US labs are lacking models in that weight class.

@maxedapps nice! how can we make it even better?

Did more testing. I'm honestly impressed by Grok 4.5. Definitely was not on my list of models to be excited about this week. Well, I guess only GPT 5.6 Sol was on that list... And Fable, of course. Anyways, Grok 4.5 is really, really good! At implementing (it's fast) but also does a good job at reviews and planning. I'm bouncing back and forth between GPT 5.5 (I'm a peasant, don't have 5.6 yet), Fable 5 (for the more complex tasks and plans), and Grok 4.5. Much, much fun!