lumost

Then why does the expensive treatment cost 7 figures? Estimating hospital costs is hard, but a quick glance at many buildings/offices would tell you that they are by no means struggling.

We overpay for healthcare because we must pay, and it has institutional and natural bottlenecks which restrict availability and choice.

A lot of big enterprises struggle with ai adoption. Either they can't get models/tools approved fast enough, can't provision them in their internal network, or get slammed with exorbitant inference costs.

A big enterprise can drop one (or 4) of these on someone's desk and let them go nuts.

There are many enterprise environments where running modern models is... difficult. Rather than fighting security for months, a user could petition for a mac studio and have rough cost parity with a z.ai subscription.

if they are a heavy user, perhaps they string 4x together.

As a kick-starter project is this thing actually real? My experience with kick-starter has been that the projects are almost universally faked.

The answer is likely that it makes sense for maintenance purposes, but the legal authority granted to an HOA is overly broad.

In the past, a professional job might mean extended periods of travel by rail etc with little work to do. A small acreage farmer has to wait for plants to grow.

The advent of factories with limitless and constant demand for labor is a new development of the industrial revolution.

Token cost is falling rapidly for a given quality. We don't think about this too much as newer models have made legacy apps obsolete - but its an interesting question what the cheapest text to sql or similar model would be. I still go by 10x cost decrease per 6 months for any given model quality.

betting on a compute bottleneck sounds like a recipe to get thrashed when the bottleneck relieves itself.

At the investment scales being discussed, CUDA/architecture and other advantages do not matter - you could spend 1 billion on building a new chip architecture. The ram/fab inputs have been a commodity market for years. Heck, even the model bottleneck doesn't seem real when it's only 1-4 billion or less to get a state of the art model.

At some point the compute bottleneck will be relieved, you can see NVidia hedging their strategy with both open models and on-device chips targeted for local inference. The 200 dollar a month plan will absolutely be taken over by local hardware in the future.

why would anybody believe true brilliance? They would only believe something that can be proved from combining existing knowledge.

If a time traveler went back to 1600 and started spouting off about differential equations everyone would think them quite mad.

I'm still uncertain if mythos is real. Subsequent model releases have been lackluster, no one has claimed to verify mythos performance and it's silently vanished from most comparisons.

I would presume the goal is to have an electromagnetic accelerator of inbound charged ions to serve as thrust. You can get limitless power from solar, but propellent mass is a limiter.

It seems like a data/roadmap acquisition. Reports have been that Cursor's footprint was still growing rapidly, and their composer model wasn't bad.

Take that data / training recipe and scale it up with compute to see what happens?

I think you'll find that 95% of all academic presentations are telling stories out of other peoples work.

There is no requirement to own the hardware. Even if you own your servers and GPUs - you would not own the power plant, real estate, or internet infrastructure to support it.

It's trivial to rent all of the above components in a competitive market under different periods, you may simply rent the token output from someone who put in the effort on this as well.

An open frontier model induces margin compression on training and inference prices across the industry.

Governments may be uniquely positioned to offer housing incentives as they often own the land, and control the zoning.

Interesting, this presents an meta for an early stage investor. Fund a company with a BS forecast based on your ability to make money on the claw back provisions.

if someone says they have a billion users because they are connected to the doubleclick exchange... they aren't lying, they might be deceptive or simply foolish.

As part of due diligence, the buyer/investor should ask how these numbers are calculated and make their own judgement. Unfortunately I believe startups select for those who dance the border of deceptive and foolish.

I think we have unfortunately apt historical comparisons. The economic state of Eastern Europe at the turn of the 20th century matches a society where wealth had been completely and utterly bifurcated.

The result was “seizing the means of production”

It's been discussed for decades that the US would run out of precision munitions in any multi-week conflict. The presumption was that the US would either be in a nuclear war at that point, or a ground campaign would complete the task along with dumb munitions.

That the US decided to expend it's precision weapon stockpile on a conflict of choice with no ground or dumb bombing campaign to fallback on is a mystery for later generations. It is not unbelievable that Iran is capable of blocking a dumb bombing campaign indefinitely, or that the viable targets for such a campaign are politically untenable.

how do you optimally balance code generation and virtual function calling patterns for the best performance? Is there a standardized OLTP format on object storage? How do you performantly implement differential privacy policies within the db engine?

Are there re-usable query primitives for extremely large scale multi-modal data? how do you scale such queries or make them efficient?

The list goes on.

I'd ascribe "deep tech" to anything that you can reasonably get a PhD in and have it not be unusual. There are dozens of academic conferences on DBs pushing the frontier forward.

Deep tech fields have a more complex relationship with academia than other industries. Engineers often need license to test ideas that may not see adoption for years, or require tens of millions to build-out.

A "research" arm is exactly this license, although it comes at the cost of potentially killing innovation in the rest of the company.

If you are asking a human to review something, it should have been verified/reviewed by a human first.

I have no interest reading someone else’s ai output that has not been verified.

There will be cloud/SaaS vendors who have lower cost of labor/capital due to automation and financing terms.

Having these models in the open caps the inference margin.

We also may be at the wealth inequality level where prices become… weird.

If there is really only a few dozen people doing the buying and the selling at the top on a weighted basis, then the prices are whatever they convince themselves of.

These companies have valuations reflecting a debt light business. At a minimum, 420 billion in debt is enough to change the stock price by 10-20%. If the company plans to add another 400 billion in debt you need to give it the side eye.

If 50 billion in revenue is from other companies debt spending… then You have a problem.

yes, above 50% I don't observe a significant improvement, the models scoring above 65% like opus4.8 seem worse than gpt5.5/opus4.6. DeepSWE appears to align more closely to my personal experience.

Does this even look like a timing race at this point? Moonshot only really got going this year.

Why couldn’t any well capitalized firm get to a leading position?

Honestly my anecdotal experience is that fable is benchmaxxed. I have not observed useful gains for Claude since 4.6, with each model iteration making progressively poorer decisions in pursuit of its goal.

The 5.5/5.6 series has performed quite well however. My guess is that my use cases stop aligning to swebench pro around 50% accuracy, and more closely align with DeepSWE.

The nature of ML is to reach for easily measured proxy problems which can then be optimized. Why worry about a backoffice integrated micro-service you have no data on when you aren't sure that you can design and implement a rewrite of an OSS project?