ypcx

Writing this from my 2019 XPS 15 7590 i9 with (extended) 64 GB RAM. The OLED display quality is still the best any laptop can offer, but that's where it ends. Had countless issues with sleep, reboots after being "killed in sleep" for going over Windows sleep "battery drain limit" (can be increased in registry but hey), wifi/bt issues after wake-up, short battery endurance, fan noise (mitigated by undervolting w/ ThrottleStop). (In case I may sound too dramatic, check their forums: https://www.dell.com/community/XPS/bd-p/XPS)

Then there's the classic Windows Laptop "goodness" like bad keyboards, bad trackpad (too much scrolling, jagged scrolling, bad trackpad surface), necessity to emulate Linux/Git/Wsl/Cygwin for any real dev work, having elaborate install.txt procedures for setting up a new laptop.

I wanted to have a "beefy" machine but since I can now do my play things on Google Colab (and my work on company's Macbook), I just ordered the last year's M1 Macbook Air and I'm done with this Dante's Inferno of Windows Laptop ecosystem. Windows 11 just reinforced my disbelief that Microsoft can produce usable operating system before our civilization reaches Singularity.

Yep, "SEO" killed the internet. As have "Accept Cookies" prompts, "Do you want notifications" prompts, "Subscribe to our newsletter" prompts and "Why are you leaving our page" prompts. Content farms, click-bot-like farms, fake-review farms. The free internet is dead.

> That doesn't mean it "doesn't work"

If our education fails to teach, then our education has failed.

In the "Definitive Guide to Embeddings", in the figure "An illustration of One Hot Encoding", the "One Hot Encoding" table doesn't make any sense whatsoever. Am I wrong?

Learn a NN model incrementally on INSERT and then use abstraction and inference on SELECT.

Here's a gotcha for you: https://github.com/denoland/deno/issues/4568

There's a few: https://github.com/denoland/deno/issues?q=is%3Aissue+is%3Aop...

As much as I'd love to have alternatives in this particular Simulation, there are none. You either use Apple and minimize your suffering, or you don't use Apple and go live in hell, provided by companies running their product design & development by teams of chimpanzees (obviously & unfortunately).

> I really want to like Snopes

Herein lies the archetypical problem with the vanilla HN (or should I say "mainstream") crowd conformist mentality. As a human herd animal, we want to belong, but it's not okay to forgo our critical thinking to do so. Downvote away.

And just like that, we've been set back by 10 years in waiting for a chair at the Milky Way Galactic Federation of Planets, which requires a 100 year continued presence of humans in space.

Not sure why I'm always hoping that they'd depart from the dependency on the past Windows architectures and produced a modern operating system. It's not that they don't know how, their labs had some very interesting concepts over the years.

The imagination then continues -- the users of this cryptocurrency-based break-away economy then establish a voluntary tax system, where each user sends crypto to the address/wallet belonging to the respective part of the government they want to see stronger/they agree with their direction -- e.g. firemen, army, roads. The address of an entity e.g. army could further be separated by policies e.g. defence vs foreign missions. It would also be customary to publicly advertise one's own tax contributions source address (which then can be easily expanded to all addresses this address has contributed to). In other words, for the first time in history, we'd have an actual democracy.

The technology is perfectly sound and in its original blueprint it was never meant to be "traded" or denominated in an actual mainstream or fiat or government currency. So if centralized elements (i.e. major exchanges) are the cancer, then regulations such as these are the scalpel.

If adopted, a cryptocurrency can perfectly serve an ecosystem as an exchange of value, without ever crossing the barrier to the mainstream monetary systems. I.e. a loaf of bread for a certain fraction of a Bitcoin.

That may depend (or be influenced by) on whether the results are negative or positive versus unvaccinated and who advertises in said journal.

Imagine being on your death bed with your only contribution to this world having been vaguely worded abstract patents which didn't translate to any real products. Especially knowing that had you not helped to stifle medical/tech innovation, maybe you would still have a few extra centuries to live. Yep, that must feel like being a cockroach and a failure in one.

Also, the article has a wrong title. A patent office would never undermine their raison d'être. In this case they're only asking to be compared to other countries so that they don't issue fewer patents than them.

The patent system is so bad that, at this point, the only way to fix it is to completely abolish it. Do you still believe that the patents are here to protect you, a garage inventor, against the mega-corps with unlimited free-money bank funding?

The real philosophical question here is where do your thoughts come from - did you really create them, are you really so special that given enough time nobody else would come up with that idea? Are you aware of the concept of multiple discovery?

  https://en.wikipedia.org/wiki/Multiple_discovery
  https://en.wikipedia.org/wiki/List_of_multiple_discoveries

On the Apple TV 2021 4K's remote you can disable the touch and just use the buttons, like I have done. Works a treat.

Cholesterol/saturated fat will clog your penis arteries which is the leading cause of ED. (In fact ED is one of the first symptoms of heart disease, the No.1 killer.)

Nutrition Facts has a page about testosterone, I'd suggest you start there: https://nutritionfacts.org/2017/05/16/foods-that-affect-test...

edit: link to a video about ED vs diet: https://youtu.be/OVXzT_RBxFk

> "Counterfeiting and inflation are the same thing. Counterfeiting is criminalized inflation, and inflation is legalized counterfeiting."

> Yes, and it’s vastly less than Bitcoin.

Ponder, for a moment, the way in which governments use the "printing" of fiat money to fund wars (for private sector profit, revolving back into that private sector's lobbying those governments for even more war and military spending).

Ponder, for a moment, the amount of energy and Earth's resources expended to materialize these wars and military spending.

Bitcoin (or, the proof-of-work cryptocurrency concept as such) by its design prevents (or at least very severely inhibits) this type of governmental hijacking for a mere fraction of the resources spent.

This applies to other heavily-subsidized sectors/areas/industries with a powerful lobbying influence/revolving doors/etc.

Imagine if we could vote by paying our taxes in Bitcoin (in a Bitcoin full-blown, mainstream Bitcoin economy) in proportion to specific Bitcoin addresses/accounts allocated for specific policy issues or target governmental groups, and thus have an actual, real and direct democracy instead of a 4-year theatre of false puppets and promises.

It's possible that they will remove the Touch Bar because the main screen will get a touch support. The Touch Bar API could then be migrated to a specific region on the main screen, without app devs having to do any extra migration work.

Add this to your browser's search engines:

  Stack Overflow
  so
  https://www.google.com/search?q=%s (site:stackoverflow.com OR site:superuser.com OR site:stackexchange.com)

Then just begin the search by typing: so[spacebar]<your search term>

"and my Fastmail account is now configured to enable me to send using my Gmail address from directly within Fastmail"

I don't see this on their website, will this keep working if Google bans your account?

I was trying to lose weight with calorie restriction, keto, intermittent fasting, etc. but it had never really fully worked. I wish I had known earlier that even a 7 day prolonged water-only fast can almost completely get me rid of any excess weight, and make me much healthier overall. For more info, check either Loren Lockman (Tanglewood Wellness Center, Costa Rica) or Alan Goldhammer (TrueNorth Health Center, California) - they both have a nice YouTube presence.

A few more transformer implementations that I’ve found:

https://github.com/pbloem/former/blob/master/former/transfor...

https://github.com/openai/blocksparse/blob/master/examples/t...

https://github.com/google/trax/blob/master/trax/models/trans...

Can't edit the post anymore so adding it here - further research reading on improving the current attention model: https://www.reddit.com/r/MachineLearning/comments/hxvts0/d_b...

I think the question of "how much can it reason" can be made more specific as "how far away can it reason from the learned examples".

Increases in reasoning power should allow for much smaller usable models.

Yes this seems like an early work in progress, compared to Jay's previous Transformer articles.

In addition to your link, I've found a really good Transformer explanation here (backed by a Github repo w/ lively Issues talk): http://www.peterbloem.nl/blog/transformers

Additionally, there's a paper on visualizing self-attention: https://arxiv.org/pdf/1904.02679.pdf

A paper on arXiv concludes that "The results suggest that limited access to publications may limit some scientific research from achieving its full impact".

https://arxiv.org/abs/2006.14979

It's not a matter of belief at this time. The evidence is right in front of our eyes. But it takes intelligence to recognize intelligence. Contextual extension by abstract inference is the basic (and hardest to achieve) building block of AGI, and we have achieved it. The rest is about utilizing this same power for the querying (priming) part of the reasoning loop, within the contexts we are interested in.

Part of me is glad we're not exactly there yet, because the thought of this running autonomously in a thinking cycle is downright scary. What will you find when you sit behind the console in the morning? It took us months to start understanding this in its current one-shot mode.

I don't care about the ideological downvotes, but we will do better if we start taking this very seriously. It's no longer theoretical that this (and machine learning as such) will have unprecedented (and impossible to predict) impact on everything we know, and the timeline is now measured in months instead of years or decades.

To ignore people who are clearly below your level of intelligence is also a sort of intelligence test that the Universe hands to you. Don't fail it.