Recurecur

So, I was downvoted twice, yet neither of you doubtless brilliant individuals bothered to refute my points.

I’ll emphasize my main point further, LLMs have no ability to innovate much beyond the vast data they’ve scraped. They’re almost entirely derivative.

They are a definite improvement over traditional search engines though!

The expansion of the universe is largely due to the impulse provided by the Big Bang. General Relativity does attempt to explain some details regarding what’s happened after the Big Bang, but provides no insight into the Big Bang itself. Nor does it offer insight into Dark Energy, a concept that Einstein opposed.

In my opinion, LLMs provide one piece of AGI. The only intelligence I’ve directly experienced is my own. I don’t consciously plan what I’m saying (or writing right now).

Instead, a subconscious process assembles the words to support my stream of consciousness. I think that LLMs are very similar, if not identical.

Stream of thought is accomplishing something superficially similar to consciousness, but without the ability to be innovative.

At any rate, until there’s an artificial human level stream of consciousness in the mix for each AI, I doubt we’ll see a group of AIs collaborating to produce a significantly improved new generation of AI hardware and software minus human involvement.

Once that does happen, the Singularity is at hand.

It’s well understood that the expansion of the universe is not “due to general relativity”. General relativity does explain some details of that expansion.

FTL starships in an SF story don’t need a detailed explanation, just a new invention.

It’s the exact same thing as a speculative story in the 1920s discussing supersonic flight, even though the jet engine hadn’t been invented yet.

For instance “Tunnel in the Sky” bypassed the whole issue in the 50’s, later imitated by “Stargate”…

Hence the movement to eliminate the “Department of Education”, which at this point would be better named the “Department of Intentional Ignorance”.

That’s despite spending more per child than any other country in the world.

All the smart and informed Americans…

Yes, COVID was a problem, along with many of the ill-considered responses to it. However, it’s now four years later, and it’s misguided to attribute the situation today to mostly being a result of COVID.

Eliminating the DOE and most of our current approach to K-12 education seems entirely prudent!

Just compare what students learned circa 1900 before graduating high school to today…

Pete Hegseth in particular is addressing those issues… :-)

That is a sad story. Now, balance it with the thousands of sad stories arising from illegal aliens killing US citizens.

That is not to mention around 100,000 US citizens dying from drug overdoses every year, due to porous borders. Thankfully, President Trump’s policies seem to be making a big difference there.

These are all simply facts.

Aloha!

They’ll be bought/leased, providing direct profit. Also, there’ll be maintenance revenue. I think they’re expected to cost around $30K.

In the case where they’re replacing a low-skill human worker, they’ll pay for themselves in 1-2 years…plus no sick days, no drug use, no theft, and they can work 24 hours a day, less any recharging time.

> Agreed. We will be screwed once China surpasses us in chip fabs, and they will. The idea that we can get a "durable advantage" by reaching AGI a few years before China is ridiculous. Using that to justify bans that only slow them down a few years at the cost of creating a chip fab juggernaut later is folly.

I’m quite sure advanced semiconductor fabs are considered a strategic necessity by China regardless of restrictions. Further, China is now getting the H200 chip…

> Tesla has rare earth free alternatives. There is no urgent need for them right now but they can switch if necessary.

There are also plenty of rare earth extraction projects coming online outside of China!

Outside the purely electric vehicles (where I believe Tesla competes very well, where is BYD at with FSD?), is there a Chinese equivalent to:

- The upcoming EREV (mostly electric extended range hybrid) F-150 truck? This is expected to have ~700 mile range, and of course no charging hassles. It’s main advantage over the now defunct Lightning will be towing range.

- The Chevy Corvette Stingray? Say what you want, but the high end ICE sports cars have an appeal of their own…

I believe the USA still has an edge in some areas of the market.

The concerns expressed seem a bit silly, unless the various Euro systems didn't take the very basic approach of using open standards and avoiding lock-in. Oh, and they should be backing up their data somewhere besides "in the cloud".

If those very basic precautions had been taken, migrating to a Euro cloud, or a private environment (open cloud stack) would be trivial.

If not, a lot of people should be fired...but granted, there are a lot of stupid people out there...

All that said, I'd say the concerns around this are vastly overblown.

We certainly do need nuclear to replace coal.

Coal and nuclear are both reliable sources of energy, unlike solar or wind. I'm sure you're about to go into some battery related hand waving, but that doesn't help in two ways. First, there will be intervals where the lull in production exceeds storage capacity, which means fatalities in many cases. Second, battery is expensive, so the total system cost can exceed the cost of nuclear (especially with sane regulation and continually improving engineering).

Also, as far as "coal trending downward", that isn't the case outside the US.

We're going to need a vastly increased energy supply in order to meet the world's needs. Nuclear fission absolutely needs to be part of that mix, as we work to master fusion and other advanced, safe, and environmentally responsible forms of energy production.

Beyond that, high density energy sources are highly desirable off-planet. :-)

Um, the DOGE is rather a necessity about now.

The USA is ~$36 TRILLION in debt. Interest payments are now consuming ~20% of tax revenue...more than the defense budget. That's insane!

Musk is also absolutely correct about over-regulation and government bureaucracy stifling innovation and economic growth.

Then there's the amazing incompetence often displayed by the current admin... Witness Afghanistan, among many other examples.

Given a two party system, I can only think this election outcome was better than the alternative!

Ha, someone on my project does exactly that...it never occurred to me it might be for perceived speed, I just figured ADHD... :^)

One major problem with bureaucracy is that it will create issues to justify itself.

I'm sure there's no real problem, it's just the government getting some PR and potentially some money via some kind of license/permit fee.

Musk and SpaceX have no desire to hurt anyone, that would impact their plans.

At this point this country would greatly benefit from less government involvement in just about everything. /libertarian

This is a silly take, primarily because it's centered on a single benchmark, Lightroom image import.

Image import is generally I/O bound, so not a good fit for CPU comparisons.

The GPU issue is relevant, we'll see how the M2 does there. Will Apple need a discrete GPU to compete?

All that said, the M1 is most impressive in terms of performance/Watt. We'll see how the M2 holds up against Threadripper/Epyc once the Mac Pro refresh is done and it's benchmarked with many CPU/GPU bound pro workloads.

The M2 should be nice for Macbook Pros though. Looking forward to it!

>One would say the target audience of Julia - matlab converts, would barely use that feature (macros) directly.

The 'target audience of Julia' is everyone, not just 'Matlab converts'.

It is most focused on high performance numerics, but is a true general purpose programming language. I expect its use to grow explosively as more and more people recognize its elegance and power.

My perspective is that of a senior engineer with decades of experience with dozens of programming languages.

Nice writeup! I think Taiwan will remain free for at least a few years, as China would pay a very heavy cost for not too much gain invading Taiwan in the near future. We'll see if Taiwan goes nuclear in the interim. The Hong Kong experience might be pushing things in that direction.

I completely agree about space being a viable new frontier. People don't seem to get that with effectively unlimited energy, habitats throughout the Solar System and even out to the Oort cloud can be arbitrarily luxurious and enjoyable.

Advances in spacecraft design and propulsion will only accelerate things... heh

Progress is sadly not instant.

Julia seems to be gathering momentum rather than losing it. The future looks bright if trends continue!

(As to the startup time problem, significant progress is being made... Have you looked at standalone binaries?)

> You can have all those wrapped under a high level API, with much nicer implementations, with more support, more documentation, more eyeballs (due to more adoption) in a "slow" language.

You 'can' have those things, but 'should' you? Julia is well designed, so 'nicer' is not a given. Julia is also quite a high level language with excellent metaprogramming facilities. Otherwise, everything you listed is based on adoption level.

Why use a "slow" language if you don't have to? Especially when the faster language is actually better designed...

'Also it is not object oriented which is probably the most popular paradigm.'

Popularity isn't a suitable metric for 'good'. Look at McDonald's food.

'Python will not be overtaken soon by such an exotic language.'

Python and Julia aren't competitors. Python is suitable for scripts and gluing high performance code together. Julia is a general purpose language which is suitable for writing high performance code. Object orientation was likely a mistake. The industry is slowly moving away from it. There is no abstraction you can express with object orientation that you can't express with multiple dispatch. Structs are very much like the data part of objects.

Even if dormant tardigrades were delivered to the lunar surface, they wouldn't be viable after a short time. Between being frozen to less than -200F at night, to roasted at 200F during the day, and irradiated with all kinds of hard cosmic radiation, any DNA present would be destroyed fairly quickly.

It's surprising to me that tardigrades can survive even a brief exposure to space.

Thanks, I hadn't run across that writeup, though well aware of the issue.

This was a little gem in that article, seemingly highly relevant now:

> However, Rawls qualifies this with the assertion that under extraordinary circumstances in which constitutional safeguards do not suffice to ensure the security of the tolerant and the institutions of liberty, tolerant society has a reasonable right of self-preservation against acts of intolerance that would limit the liberty of others under a just constitution, and this supersedes the principle of tolerance. This should be done, however, only to preserve equal liberty – i.e., the liberties of the intolerant should be limited only insofar as they demonstrably limit the liberties of others: "While an intolerant sect does not itself have title to complain of intolerance, its freedom should be restricted only when the tolerant sincerely and with reason believe that their own security and that of the institutions of liberty are in danger."

This isn't correct.

A fair number of these benchmarks were multithreaded, and the M1 held its own very well, despite being at a substantial disadvantage (8 threads (only 4 on performance cores) for M1 versus 24 for Ryzen 3900X).

That said, single threaded performance is extremely important, since not everything can be parallelized. Amdahl's Law, eh?

Don't worry though, I'm sure Apple Silicon is coming with many more cores. Amazon's Graviton2 shows how ARM can perform with massive core counts...

As to the current Ryzen (5000 series), it's a great chip and what I'm planning to use in a new build, but it also has a 105W power budget. We'll see what Apple produces for its desktop entries soon enough!

The grandparent used the wrong term. The memory isn't on the die, but it is in the package (MCM). I don't think that provides faster throughput, but it should help latency.

Regardless, the big win was giving unified access to that memory using a memory fabric. That lets the GPU (for instance) access graphics assets with no copying - it just uses them in place.

AMD can absolutely do the same thing, if it desires. Intel can too, but it has a hill to climb WRT GPU performance.

Apple's AI chiplet is another matter...

Battery life alone isn't the main point!

Extreme battery life + extreme performance is the main point!

The Thinkpads with extra battery weren't in the same "small and light" class as the Air and MBP 13. These "small and light" laptops with 20 hour video playback are faster than current workstation class laptops!

> But claiming that what’s available today from Apple is of great interest to the high performance market is jumping the gun.

I think that misses the point that most users (and most software developers in particular) use laptops. The M1 SOC pretty well kills everything CPU-wise, and does graphics as well as a GT 1050 discrete GPU. It also throws in AI silicon and RAM, and runs all of it in a 20W (or so) power budget.

Single-thread performance is quite comparable to a Threadripper.

It'll be interesting to see how things look when M1 style silicon is introduced for the 16" Macbook Pro (guessing 8 performance cores, 32 GB RAM), and the Mac Pro (guessing up to 32 performance cores, and at least 64 GB RAM used as cache for a potentially much larger main memory).

> There’s also the fact that cost (including peripherals and support) and software availability, not performance and efficiency, are what guide most PC purchases. The M1’s advantages are far less pronounced in those areas.

Actually as to cost, it's highly competitive in the premium laptop segment (like Dell XPS for instance).

Software availability is good, since it'll run almost all Intel Mac software, all M1 Mac software (soon to be ~everything), and all ARM-compatible Linux software (ARM Linux runs in a VM).

The only thing missing is Windows, and that may well be temporary.

I'd say if you're doing web, Mac or Linux software development the Apple Silicon Macs are the wave of the future! :-)