when you want to use Google, you go into Google search, and you type in English, and it matches the English with the English. What if we could do this in FreeSpeech instead? I have a suspicion that if we did this, we'd find that algorithms like searching, like retrieval, all of these things, are much simpler and also more effective, because they don't process the data structure of speech. Instead they're processing the data structure of thought

why is it that we’re not focusing on those movements as the source of our strength and our organizing? It’s because we have a discourse framed around elite policy institutions that make them the primary actors and the coordination of mostly market mechanisms

The fossil fuelindustry is a significant contributor to the Big Green organizations, and many of theseorganizations are financially invested in renewables and fossil fuels, so they do not want tosee nuclear power as a competitor.

Cloud Capital Silicon Valley in the United States and finance don't work together Apple pay exists Google pay exists so like wej you can pay through Apple pay but a large segment of that money goes to Wall Street as a rent so there is a class like something like class war between or a fudal l war between the fom of Wall Street on the West CO east coast and the fom of cloud capital on the west coast they're clashing that Clash doesn't happen in China because both your Finance sector and your big Tech or Cloud Capital sector are under the Communist party

I would say the epigenetic inheritance that has to occur there and how it occurs must be contributing a very large fraction indeed to the differentiation process

in the data center you're dealing with things at the microsc or millisecond scale uh when you move out to the edges of the network you're dealing with seconds and minutes

that's part of the logic of agriculture isn't it i mean you have a lot of work to do yeah a lot of people you know but there again you're you're in a progress trap or or a treadmill that you need more children 00:48:57 so you can work more land and then that more land provides more food so you have yet more children

deluding the general royalty pool

LLMs, by their very nature, don't have a concept of "source". Attribution is pretty much impossible. Attribution only really works if you use language models as "search engine". The moment you start generating output, the source is lost.

GPT-4 System CardOpenAIMarch 23, 2023

Training language models to follow instructionswith human feedback

LaMDA: Language Models for Dialog Application

It was only by building an additional AI-powered safety mechanism that OpenAI would be able to rein in that harm, producing a chatbot suitable for everyday use.

It’s surprising because these models supposedly have one directive: to accept a string of text as input and predict what comes next, over and over, based purely on statistics. Computer scientists anticipated that scaling up would boost performance on known tasks, but they didn’t expect the models to suddenly handle so many new, unpredictable ones.

Dass das ägyptische Wort p.t (sprich: pet) "Himmel" bedeutet, lernt jeder Ägyptologiestudent im ersten Semester. Die Belegsammlung im Archiv des Wörterbuches umfaßt ca. 6.000 Belegzettel. In der Ordnung dieses Materials erfährt man nun, dass der ägyptische Himmel Tore und Wege hat, Gewässer und Ufer, Seiten, Stützen und Kapellen. Damit wird greifbar, dass der Ägypter bei dem Wort "Himmel" an etwas vollkommen anderes dachte als der moderne westliche Mensch, an einen mythischen Raum nämlich, in dem Götter und Totengeister weilen. In der lexikographischen Auswertung eines so umfassenden Materials geht es also um weit mehr als darum, die Grundbedeutung eines banalen Wortes zu ermitteln. Hier entfaltet sich ein Ausschnitt des ägyptischen Weltbildes in seinem Reichtum und in seiner Fremdheit; und naturgemäß sind es gerade die häufigen Wörter, die Schlüsselbegriffe der pharaonischen Kultur bezeichnen. Das verbreitete Mißverständnis, das Häufige sei uninteressant, stellt die Dinge also gerade auf den Kopf.

"There is a robust debate going on in the computing industry about how to create it, and whether it can even be created at all."

For example, when an AI technology receives solely a prompt [27] from a human and produces complex written, visual, or musical works in response, the “traditional elements of authorship” are determined and executed by the technology—not the human user.

L.L.M.s have a disturbing propensity to just make things up out of nowhere. (The technical term for this, among deep-learning experts, is ‘‘hallucinating.’’)

More interesting or alarming or hilarious, depending on the interlocutor, is its propensity to challenge or even chastise its users, and to answer, in often emotional language, questions about itself.

Scaling a single VCS to hundreds of developers, hundreds of millions lines of code, and a rapid rate of submissions is a monumental task. Twitter’s monorepo roll-out about 5 years ago (based on git) was one of the biggest software engineering boondoggles I have ever witnessed in my career. Running simple commands such as git status would take minutes. If an individual clone got too far behind, it took hours to catch up (for a time there was even a practice of shipping hard drives to remote employees with a recent clone to start out with). I bring this up not specifically to make fun of Twitter engineering, but to illustrate how hard this problem is. I’m told that 5 years later, the performance of Twitter’s monorepo is still not what the developer tooling team there would like, and not for lack of trying.

In very large code bases, it is likely impossible to make a change to a fundamental API and get it code reviewed by every affected team before merge conflicts force the process to start over again.

One of the most well-documented shortcomings of large language models is that they can hallucinate. Because these models have no direct knowledge of the physical world, they're prone to conjuring up facts out of thin air. They often completely invent details about a subject, even when provided a great deal of context.

The application is powered by LaMDA, one of the latest generation of large language models. At its core, LaMDA is a simple machine — it's trained to predict the most likely next word given a textual prompt. But because the model is so large and has been trained on a massive amount of text, it's able to learn higher-level concepts.

That's greater than taking all the humans who lived throughout time, multiplied by the number of grains of sand on Earth, multiplied by the number of atoms in the universe.

LLMs are generative math-ematical models of the statistical distributionof tokens in the vast public corpus of human-generated text, where the tokens in question in-clude words, parts of words, or individual char-acters including punctuation marks. They aregenerative because we can sample from them,which means we can ask them questions. Butthe questions are of the following very specifickind. “Here’s a fragment of text. Tell me howthis fragment might go on. According to yourmodel of the statistics of human language, whatwords are likely to come next?”

The breakthroughs are all underpinned by a new class of AI models that are more flexible and powerful than anything that has come before. Because they were first used for language tasks like answering questions and writing essays, they’re often known as large language models (LLMs). OpenAI’s GPT3, Google’s BERT, and so on are all LLMs. But these models are extremely flexible and adaptable. The same mathematical structures have been so useful in computer vision, biology, and more that some researchers have taken to calling them "foundation models" to better articulate their role in modern AI.

f we can't get food services to them, it becomes easier to break those large cities up into smaller communities that are more decentralized.

Large Group Discussions

Epics, issues, requirements, and others all have similar but just subtle enough differences in common interactions that the user needs to hold a complicated mental model of how they each behave.

Nevertheless, designers can limit the value of attacks by limiting the scope of what governance can do

Perhaps not a good idea, in general, to use a random PPA for such sprawling software as a browser. Auditability near zero even if it is open source.

Such scaled-up communication and collaboration processes would also require meta-design principles to collaboratively construct the required design rationale, media and environments [23].

Etzioni astutely observed that all communities have a serious defect: they exclude. To prevent communities from over-excluding, they should be able to maintain some limitations on membership, yet at the same time greatly restrict the criteria that communities may use to enforce such exclusivity. He therefore proposed the idea of “megalogues”: society-wide dialogues that link many community dialogues into one, often nation-wide conversation [7].

When you look inside a node_modules directory, there’s likely hundreds if not thousands of packages, even for a relatively basic application.

How are hundreds of dependencies and 28,000 files for a blank project template anything but overly complicated and insane?

The incident demonstrates a type of power that Amazon wields almost uniquely because so many companies rely on it to deliver computing and data storage.

See also BMW and Tesla owners. If Tesla does become the largest US-based carmaker, many of the buyers will, I'm sure, think of reasons to move onto something else.

definite good news, as it will hopefully have a ripple effect on crappy chipset makers, getting them to design and test their hardware with Linux properly, for fear of losing all potential business from Lenovo.

I suppose it means 2 things, first, you get official support and warranty, and second, the distros will be Secure Boot approved in the UEFI, instead of distro makers having to figuratively ask Microsoft for pretty please permission.

If you want to implement a form with a superb User Experience, you have to take care of many variables:

One of the primary tasks of engineers is to minimize complexity. JSX changes such a fundamental part (syntax and semantics of the language) that the complexity bubbles up to everything it touches. Pretty much every pipeline tool I've had to work with has become far more complex than necessary because of JSX. It affects AST parsers, it affects linters, it affects code coverage, it affects build systems. That tons and tons of additional code that I now need to wade through and mentally parse and ignore whenever I need to debug or want to contribute to a library that adds JSX support.

I forgot to mention in the original issue way back that I have a lot of data. Like 1 to 3 MB that is being passed around via export let foo.

I just created an empty Rails project and just find node_modules -print | wc -l gives me 18366 files!

Since re-rendering in Svelte happens at a more granular level than the component, there is no artificial pressure to create smaller components than would be naturally desirable, and in fact (because one-component-per-file) there is pressure in the opposite direction. As such, large components are not uncommon.

In my projects on Svelte, we adhere to the "budget" of the component in 200 loc with styles. If the component goes these limits, we just take out styles in a separate file using svelte-preprocess.

Eliminating needless classes frees up operating memory on the computer—and bandwidth in your head.

Your styles are scoped to the component. No more leakage, no more unpredictable cascade.

It's fashionable to dislike CSS. There are lots of reasons why that's the case, but it boils down to this: CSS is unpredictable. If you've never had the experience of tweaking a style rule and accidentally breaking some layout that you thought was completely unrelated — usually when you're trying to ship — then you're either new at this or you're a much better programmer than the rest of us.

It gets worse when you're working on a team. No-one dares touch styles authored by someone else, because it's often unclear what they're doing, what markup they apply to, and what disasters will unfold if you remove them. The consequence of all this is the append-only stylesheet. There's no way of knowing which code can safely be removed, so it's common to undo some existing style with another, more specific style — even on relatively small projects.

Some large tech behemoths could hypothetically shoulder the enormous financial burden of handling hundreds of new lawsuits if they suddenly became responsible for the random things their users say, but it would not be possible for a small nonprofit like Signal to continue to operate within the United States. Tech companies and organizations may be forced to relocate, and new startups may choose to begin in other countries instead.

Other sites could absolutely spend time crawling for new lists of breached passwords and then hashing and comparing against their own. However this is an intensive process and I'm sure both Facebook and Google have a team dedicated to account security with functions like this.

Ultimately it comes down to how much time and money you can dedicate to keeping your users' accounts secure versus how important it is to do so. Google and Facebook accounts sit at the centre of many users' internet lives and would be devastating to use. Same for most email accounts.

Patients with persistent pneumothorax, large air leaks after tube thoracostomy, or difficulty ventilating should undergo fiber-optic bronchoscopy to exclude a tracheobronchial injury or presence of a foreign body.

Furthermore, one should also consider that **publishers – a category including natural persons and SMEs – are often the “weaker” party in this context.** Conversely, third parties are usually large companies of substantial economic import that work as a rule with several publishers, so that one publisher may often have to do with a considerable number of third parties.

For large-scale software systems, Van Roy believes we need to embrace a self-sufficient style of system design in which systems become self-configuring, healing, adapting, etc.. The system has components as first class entities (specified by closures), that can be manipulated through higher-order programming. Components communicate through message-passing. Named state and transactions support system configuration and maintenance. On top of this, the system itself should be designed as a set of interlocking feedback loops.

In this FLT article, I am introducing a new pedagogy I call the Pedagogy of Retrieval. This is the pedagogy I use to try to interrupt the automatic use of lower potential learning strategies in my flipped classrooms at The University of Texas at Austin, and it is built on the collective body of research and efforts of my colleagues mentioned above.

Clark’s work with developmental math is part of a bigger transformation going on at Oregon State. A three-year, $515,000 initiative funded by an Association of Public and Land-grant Universities (APLU) grant is enabling educators to overhaul eight high-enrollment general education courses classrooms with adaptive and interactive learning systems.

for not very large numbers

The creation and exploitation of large-scale quantitative atlases will lead to a more precise understanding of development.

Just as comprehensive datasets of genomic sequence have revolutionalized biological discovery, large-scale quantitative measurements of gene expression and morphology will certainly be of great assistance in enabling computational embryology in the future. Such datasets will form the essential basis for systems level, computational models of molecular pathways and how gene expression concentrations and interactions alter to drive changes in cell shape, movement, connection, and differentiation. In this review, we discuss the strategies and methods used to generate such datasets.