@Romy_Holland there are hotels in the uk set up for this kind of thing, you get given a monitor on arrival with hotel-wide signal and you can go the restaurant / spa / bar etc. in the evening.
English
George Powell
94 posts
George Powell
@TheGeorgePowell
Research Engineer @ DeepMind Logic, computation, optimisation.
London Katılım Ekim 2012
205 Takip Edilen117 Takipçiler
All non halting computer programs either get stuck in a simple loop or have a memory leak.
EduardoRFS.tei@TheEduardoRFS
@KleeneAlgebra Oh, there are non-terminating programs, they're just regular recursion, aka you will see the same state repeat after a while which means if you just store the states, you know if it's non-terminating or not.
English
honestly if it wasn't for that scene from limitless, the one where he just does the dishes after gaining super human focus and energy, I'd be a much less functional adult.
English
@walterkirn Whatever their shortcomings, they're mind bogglingly good at coding and the software engineering profession is unrecognizable from 6 months ago. Most of the hyperbole will be from people who have experienced that first hand.
English
How is it that the LLMs get things wrong constantly, the very simplest things, and make stuff up pretty much nonstop, yet they are said to be hurtling unstoppably toward god-like power -- if they haven't secretly achieved it already?
Is this a con job?
English
@MegaBasedChad I was 9 years old and it was one of the most vivid memories I have from childhood. Fireworks going off everywhere while telling my mum it was the best day of my life. The way I saw it, I was experiencing something less than 1/10 generations can.
English
Question for Uncs who were there when it happened
How did the turn of the millennium feel? What was it like?
Nostalgia@NostalgiaFolder
The very first sunrise of the 2000s, marking the start of a new millennium
English
great single pitch tweet for bend. really hope this catches on.
Taelin@VictorTaelin
it is not fundamentally different, just the language is designed to software and apps first, not proof first. goals include: → absolute performance (including C/GPU compilers) → a massive stdlib with IO, file, networking, etc. → compiling to efficient web apps (like React / Elm) → a type checker that scales (10k+ files in a second!) → a syntax familiar to devs, not mathematicians among many other things the type system itself is underwhelmingly simple compared to Agda/Lean. no homotopy stuff, no quotients, not even unification, just a plain old and fast bidirectional type checker. but that's what makes it scale!
English
Do you read the types? How thorough are they? Like are all the important behaviors included in the types? How easy is it to understand and validate the types?
I am a true believer in this project and hope it succeeds but I do wonder how effective it is at reducing bugs if building the right types is still hard.
Do you have unit tests too or just types?
Sorry loads of questions!
English
exactly the point of Bend2!
if you 1. specify your program with strong types (aka theorems). 2. let the AI prove it, 3. let the compiler check it, then you *literally* don't need to see it. there is no reason to. it IS correct!
btw, Bend now has nearly 350k lines of code, 264k of which are in the stdlib. agents are autonomously porting entire libraries from Rust, Haskell, etc. to Bend, and everything is going smoothly
I barely check it since I'm working in other things
just codex /goal for now, every day it grows larger
soon, every algorithm ever conceived by humans will be in Bend
solst/ICE of Astarte@IceSolst
Interesting article on treating agent output like compiler output (and why) skiplabs.io/blog/codegen_a…
English
There's this dynamic in software engineering where someone builds a system w. good abstractions and then someone builds an easier-to-use but less flexible layer on top of that which becomes popular and then people build on top of that to shoehorn some version of the original flexibility on top again because people are afraid of working with "low level" things.
Worst case is how complicated ORMs got as a way of avoiding writing SQL.
English
@Aella_Girl It doesn't have positive valence universally though, like people wouldn't see it as good in a CEO or politician.
It also doesn't just mean exposure/experience. It can refer to uncorrupted intentions, which maybe deserves the positive valence in general.
English
3x this week
I sat in 200°F (93°C) dry sauna
for 55 min
to get my core body temp to 102.4°F (39°C)
That temp, at that duration, triggers primal panic.
Took blood samples before and after each session to measure heat shock protein expression.
Sauna is one of the world's best longevity therapies. What we don't know is whether hitting the core temp threshold unlocks a meaningfully superior response or whether the benefits are more fundamental: vascular shear stress, nitric oxide release, sweating.
Most people doing sauna never reach 102.4°F. That's high-grade fever territory.
I'm excited to see the data.
English
@QiaochuYuan @corsaren Neither are. The lunatic that sets up the experiment is killing people and all the button pressers are victims.
English
It’s reflexive, so convincing others that you are correct makes you more correct. Yapper catnip. Yapnip.
What’s crazy is I’m still learning new metagame dynamics 1 week in. For instance: apparently there is a significant contingent who thinks that the *blue* button is evil/wrong
Slazac 🇪🇺🇺🇦🇹🇼🌐@TrueSlazac
The red/blue discourse has lasted so long it's kind of absurd, why the fuck is it so persistent
English
@allTheYud @wolflovesmelon > human extinction is a bipartisan issue
> I don't consider you on my side
🙄
English
@wolflovesmelon Hi, my name is Eliezer Yudkowsky. Guido doesn't like me, and I don't consider him to be on my side.
Human extinction is a bipartisan issue if literally anything on Earth is a bipartisan issue. Attaching it to an attack on Republicans is stupid.
May we all have nice days.
English
Hi my name is Guido Reichstadter & I'm currently occupying the top of the Frederick Douglass memorial bridge in Washington DC.
I'm calling on the people of the United States to bring an immediate end to the Trump regime's illegal war on Iran and the removal of the regime power through mass nonviolent direct action and non-cooperation.
I also want to urgently warn the people of the US and the world of the imminent danger we are in of crossing a point of no return towards the development of artificial intelligence which poses the risk of catastrophic harm to humanity, including human extinction. I call on the governments of the world to take immediate action to end this danger by permanently banning the development of artificial general intelligence and machine super intelligence. I also call on the people of the world to exert all possible influence through nonviolent action to compel their governments to end this danger with all possible speed.
English
Causal decision theory let's you get the £1m, you just need to decide to pick B early on in life before the alien predictor has finished modeling you.
If that's considered cheating, then Functional Decision Theory doesn't help since for it to work you also have to have chosen to adopt the framework before the predictor has finished modeling you.
In that sense, to the extent that Functional Decision Theory is correct, it can be seen as a logical consequence of causal decision theory when applied to a world where rewriting your own decision making process is considered an option.
DannyCanTalk 🌈@dannycantalk
We're done rehashing the button question. Time to rehash Newcomb's Paradox. Are you a one-boxer or a two-boxer?
English
Impressions from a brit:
Korean: looks like a film, probably fiction
Italian: like the British one but with worse spacing
Bulgarian: doesn't look like it comes from a serious author
Japanese: nice aesthetic but probably contains poetry I'm not interested in
Dutch, American, Portuguese: too dramatic, immediate distrust.
Spanish: probably a book about slime?
English: favorite by far
someone did a good job.
Rob Bensinger ⏹️@robbensinger
Titles of different editions of the book (according to Claude/Google): 🇺🇸 US IF ANYONE BUILDS IT, EVERYONE DIES: Why Superhuman AI Would Kill Us All 🇬🇧 UK IF ANYONE BUILDS IT, EVERYONE DIES: The Case Against Superintelligent AI 🇧🇬 Bulgaria IF ANYONE DEVELOPS IT, WE WILL ALL DIE: Why Superhuman Artificial Intelligence Could Erase Us All 🇮🇹 Italy BEFORE IT'S TOO LATE: Why Artificial Superintelligence is a Threat to Humanity 🇪🇸 Spain IF SOMEONE CREATES IT, WE WILL ALL DIE: Why Artificial Superintelligence is a Threat to Humanity 🇳🇱 Netherlands IF SOMEONE BUILDS THIS, EVERYONE WILL DIE: Why Superintelligent AI Could Mean Our End 🇧🇷 Brazil IF SOMEONE CREATES IT, EVERYONE DIES: Why Superhuman AI Can Kill Us 🇰🇷 Korea AI — GOD'S BIRTH, HUMANITY'S END: The Birth of Superintelligence, and What Will Happen After 🇯🇵 Japan IF SUPERINTELLIGENT AI IS BUILT, HUMANITY WILL GO EXTINCT Other upcoming translations include Mandarin (simplified), Russian, Ukrainian, Vietnamese, Polish, German, and Portuguese (Portugal).
English
@itaisher yeah a lot of people understand "sentience" to just be a threshold of intelligence/capability
English
I feel like this captures the conventional view among llm researchers.
Grady Booch@Grady_Booch
I am confident that the mind is computable; consciousness is an exquisite consequence of the laws of physics. I am also confident that contemporary systems such as @claudeai are without a doubt not only not conscious but are in fact architecturally incapable of consciousness. And furthermore, we are not even close to closing that gap.
English
Look guys, it's actually really straightforward, a bunch of people staked their ETH on the Ethereum blockchain to earn yield, except they didn't want their capital to be locked up, so they actually staked with a liquid staking protocol called Lido who provided them a liquid staking receipt token called stETH, except they decided to juice their yield further by depositing their stETH receipt tokens into a restaking protocol called Eigenlayer, except they didn't want to lock up their capital, so they actually restaked with a liquid restaking protocol called KelpDAO who provided them with a liquid restaking receipt token called rsETH, except they decided to juice their yield further by depositing their rsETH tokens into a lending protocol called Aave so that they could open a leveraged looping position that borrows ETH against the rsETH collateral and restakes the ETH into rsETH which is then deposited as collateral, except it turns out rsETH used a cross-chain bridge called LayerZero that was hacked by north koreans causing rsETH to become undercollateralized and now these looping positions are stuck and unprofitable, and everyone is pointing fingers at each other, and also DeFi is a very serious industry
English
what makes an addiction an addiction, structurally, is not that it feels good and you want to keep doing it - drinking water feels good and we want to keep doing it and that's not an addiction. what makes an addiction an addiction is that good feelings mean something, and sometimes those meanings are lies
when you use an addiction to constantly lie to yourself over and over again, your felt understanding of your life diverges increasingly from the reality of what's actually happening - coked out bliss while the trash piles up in your hikikomori cave. serious addictions destroy your life because they undermine the ordinary processes that cause you to wake up and go "wait, this shit is not okay, i have problems and i need to fix them"
QualiaNerd@QualiaNerd
The official concept of "addiction" is based on the axiom that it is illegitimate to want to enhance your valence technologically. You must cheerfully surrender the sovereignty over your hedonic setpoint to the FDA. If you don’t, it means you are "addicted".
English
@bitcloud wow it actually has a little face too, hard to argue against this.
English