andreyf

It's not a conspiracy to say it's unwise to expect a company to drive a research & product direction that will directly impact their reduce. Some might be smart enough to realize it's the right long game, but it takes leaders who understand and plan out second order consequences.

It's not that it fundamentally isn't able to produce something concise, it's that the business model of the companies developing these models rests on selling tokens...

No, in ~6y they will count for 1% of present day consumption. The rest of the quote compares it to where we expect steel manufacturing projections. It all reads like something that seeks to influence rather than inform.

aaronsw would be proud

Rumor has it that they weren't trained "from scratch" the was US would, i.e. Chinese labs benefitted from government "procured" IP (the US $B models) in order to train their $M models. Also understand there to be real innovation in the many-MoE architecture on top of that. Would love to hear a more technical understanding from someone who does more than repeat rumors, though.

Weekly version in sheets makes more sense for me, since almost everything in my year is split logically by week — https://docs.google.com/spreadsheets/d/18YsOpI4GsxKGZmpMRQcN...

Generated using this JS in an action script — https://gist.github.com/anfedorov/9f7dc03432a4da783577aec68f...

Not all humans, just the ones that can't afford buy property in low risk areas, start companies to help people move, etc. and for those who can, appreciating investments galore!

Great news, and hopefully more to come across other publications! If only aaronsw was here to see it :(

That's not the same thing at all. Especially as LLM's get smaller, WhatsApp can make a lot of money reading your messages and listening to your calls on the endpoint in order to market segment your ads while remaining e2e encrypted.

As a non-benefitting fan of DEI, it always seemed like too little and/or too late [1] and the current generation of efforts have paid the cost of perceived injustice that are obviously counter-productive. This change to only having diversity for corporate benefit is an understandable choice of weak leaders in the current political climate but also unwelcome by those of us who care to actually move the needle on mutual understanding and healing of past injustices. Other countries have actually done this to a great extent, I think -- healing and reconciliation on horrific historical injustices, so it's not like this is totally novel social research.

Now to some extent but even more so if things get more "rude" for a lack of a more more specific but agreed on term, I think many educated and well rounded people with a choice in employment/location will lean towards working at employers which works more aggressively towards social justice, or move abroad to live in societies not plagues by this kind of un-healing racial strife. From my understanding of what it's like to be black elsewhere (UK, Canada), I'll be more and more surprised that people who have the means would choose to stay in the US much longer.

1. a 2019 study by Darrick Hamilton and colleagues estimated that eliminating the racial wealth gap in the U.S. could require a transfer on the order of $10 trillion: https://socialequity.duke.edu/wp-content/uploads/2019/10/Run...

if someone had told me that "full of AI" was going to be an insult one day, I would have said they were too cynical about the glorious AI future, but in hindsight perhaps you are right

More likely to be a fake match so that people can't check whether records are in their database.

Was the email they offered to send the code to f**k@corpdomain.com? I tried this with my current and former employer and got seemingly the same email username "f**k", and seemingly random phone numbers. Another long-time employer had no email at all. Definitely seems they're just making things up.

Too good to be true!

yes. worth remembering also that even with e2ee, a ad-tech-driven company could have endpoints determine marketing segments based on content of conversations ad report those to the company to better target ad spend.

> You can't walk into someone's home and touch all their stuff just because they left their front door open.

There's a few big differences here: (1) the "someone" is a company with $7B/yr in revenue and dedicated security resources, also (2) with millions of users relying on their doing an Ok job securing the "home", and (3) which is connected to a network immediately accessible to anyone from bored teenagers all over the world to organized criminals and nation states looking to harm their users' interests. If a random script kiddie can find this hole by accident, how many determined teams with even minor budgets were accessing their network? Spooking KPN's security team into action was a mitzvah all around, even if he did not with whole heart and mind intend it as such [1].

More generally, the story this reminded me of most is that of Aaron Schwartz -- it's a true sadness that our societies deal so poorly with analytically developed folks' efforts being even slightly mis-applied by some fiendish letter-of-the-law measure and otherwise "good" people entrusted to exhibit moral judgement seem to go into a frothing-at-the-mouth attack against those who literally did no harm to groups that are supposedly being protected (scientific publishing / ISP users). I'm not sure if there is a theory of law by which a more-global net-good can completely outweigh more-local crimes, but it seems a society that were to allow for that would be both more successful and just.

1. https://judaism.stackexchange.com/questions/28579/do-you-get...

"Allowing somebody else to harvest it is money straight out of Facebook's pocket."

In that metaphor, Facebook owns the soil and they sell the harvest by charging for promoted posts and other forms of engagement purchasing. If this engagement truly drives election outcomes the way the post hypothesizes, the demand side will come back every time there's an election anywhere. Let's hope Zuck and co understand how closely this piece of their revenue is tied to democratic engagement ;)

don't usually flex with tech but isn't uMatrix [1] just a little bit harder to learn and a million times more satisfying? at least on my often-used systems, I have a hard time imagining loading a website without being able to control the third party content in a matrix these days.

1. https://addons.mozilla.org/en-US/firefox/addon/umatrix/

after a bunch of looking, the paper that had the most detail on quality and sensors used is https://www.sciencedirect.com/science/article/pii/S036013232...!

the most accurate readily available for CO2 seems to be the "Laser Egg+ CO2" for $200 https://shop-usa.kaiterra.com/collections/frontpage/products... which uses SenseAir S8 but looks and sounds a little silly

and the $150 "Awair Element" https://wiki.liutyi.info/display/CO2/Awair+Element+inside which uses the Amphenol Telaire T6703-5K

Those are subjective but there's a good paper on accuracy here https://www.sciencedirect.com/science/article/pii/S036013232...!

tldr for CO2 best is https://www.kaiterra.com/en/laser-egg-co2

and if you don't want to spend $250, the DIY version seems pretty easy, too: https://www.greenhousetoday.com/growing-spirulina/

Which science do you mean? Found a redacted study from Germany and doctors saying there's no effect on blood O2 or CO2 levels https://uihc.org/health-topics/do-face-masks-make-you-retain...

800ppm is "on the line" per what's industry recommended for an office http://www.energy.wsu.edu/documents/co2inbuildings.pdf

*Winix

After looking at CR and buying quite a few others I'm pretty solid on alen.com standalone units, altho a central HVAC definitely seems more effective, esp. given that you aren't necessarily recirculating a build-up of CO2 all night long.

Precisely, and in that context, let's remember that such projection is fundamental to our psychology. Thus, this story would better serve as litany of cautionary vignettes of humans being human than a story of David v Goliath, a tale of The Demanding Intern, or anything else bigger-than-life.

Universities are not top-down controlled institutions - profs have tenure (i.e. independence) and run their labs at a relatively micro scale, so whatever the criticism of tech mega-corps, it probably doesn't transfer because of those differences.

Soldiers in military often have a hard time adjusting to the freedom of civilian life, and it seems that structure is there out of institutional necessity (i.e. society could not exist without a military to protect it), not for the benefit of the individual soldier. Creating corporations that are both single-leader authoritarian and control more aspects of their employee's lives may not be how we want more of in our society. A lot of successful corporations seem relatively all-consuming for their employees (both small and big corps), so it seems like a subject without a solid path toward consensus.

Why does it sound creepy? What percentage of the population, historically, have been "free" in a sense that they don't belong to an organization that exerts some sort of control over their social world? It seems there are many shades and dimensions here...

There are immediate dependencies? Seems "depended on" should include sub-deps, as well.

Where is the shame coming from?