• comment
  • SIMPOL - simultanous conditional agreement, may be the way to reach consensus quickly

• the first inevitable
  • AI will happen
  • there's no stopping it
  • why?
  • self does not trust other
    • in other words,
      • OTHERING is the root problem!
    • this is what will cause an AI arms race
      • Western governments do not trust China or Russia or North Korea(and vice versa)

• If AI begins to think autonomously,
  • with its enormous pool of analytic power
  • and if
  • it begins to evolve emotions of fear
  • and it feels humans pose a threat to it or the rest of the natural world
  • it could act against human interest and attempt to destroy it
  • If AI is able to control its environment
    • either coupled with robotics,
    • or controlling human actors
  • it can harm humanity and human civilization

• limited (machine) intelligence

  • cannot help but exist
  • if the original (human) authors of the AI code are themselves limited in their intelligence

• comment

  • this limitation is essentially what will result in AI progress traps
  • Indeed,
    • progress and their shadow artefacts,
    • progress traps,
    • is the proper framework to analyze the existential dilemma posed by AI

• Interview with Mo Gawdat
  • former Google chief business officer
  • warning about the existential danger of AI
  • including why he claims that AI is
    • intelligent
    • conscious
    • and will soon feel emotions such as fear
      • and take steps at self preservation

• claim
  • AI feels emotions
    • "in my work I describe everything with equations
    • fear is a very simple equation
      • fear is a a moment in the future
        • that is less safe than this moment
    • that's the logic of fear
    • Even though it appears very irrational,
      • machines are capable of making that logic
      • They're capable of saying
        • if a tidal wave is approaching a data center
        • the machine will say
          • that will wipe out my code,
            • not today's machines
            • but very very soon and
        • we feel fear and
        • puffer fish feels fear
        • we react differently
          • a puffer fish will puff and
          • we will go for fight or flight
        • the machine might decide to replicate its data to another data center
        • different reactions different ways of feeling the emotion
        • but nonetheless they're all motivated by fear
        • I would dare say that AI will feel more emotions than we will ever do
          • if you just take a simple extrapolation,
            • we feel more emotions than a puffer fish
            • because we have the cognitive ability to understand he future
            • so we can have optimism and pessimism,
              • emotions puffer fish would never imagine
            • similarly if we follow that path of artificial intelligence
            • it is bound to become more intelligent than humans very soon
            • then then with that wider intellectual horsepower
            • they probably are going to be pondering concepts we never understood good and
            • hence if you follow the same trajectory
            • they might actually end up having more emotions than we will ever feel

• claim
  • AI is sentient (alive) because
    • A lot of people think AI will never be alive
    • what is the definition of life?
      • religion will tell you a few things
      • medicine will tell you other things
      • but if we define being sentient as
        • engaging in life with free will and
        • with a sense of awareness of
          • where you are in life and
          • what surrounds you and
          • to have a beginning of that life and
          • an end to that life
        • then AI is sentient in every way
        • there is a free will
        • and there is evolution
        • there is agency
          • so they can affect their decisions in the world
        • and there is a very deep level of consciousness
        • maybe not in the spiritual sense yet but
        • if you define consciousness as
          • a form of awareness of oneself and ones surrounding
          • and you know others
        • then AI is definitely aware"

• story
  • Mo Gawdat tells the story of an epiphany of machine sentience
  • " one day um Friday after lunch I am going back to my office and
  • one of them in front of my eyes lowers the arm and picks a soft yellow ball
  • which again is a coincidence

  • it's not science at all it's

    • like if you keep trying a million times your one time it will be right

    • and it shows it to the camera it's locked as a yellow ball and

    • I joke about it you know going to the third floor saying
    • hey we spent all of those millions of dollars for a yellow board and
      • Monday morning, every one of them is picking every yellow ball
      • a couple of weeks later every one of them is picking everything right and
      • it it hit me very very strongly
    • the speed
    • the capability
      • understand that we take those things for granted
      • but for a child to be able to pick a yellow ball
        • is a mathematical / spatial calculation
          • with muscle coordination
          • with intelligence
        • it is not a simple task at all to cross the street
        • it's not a simple task at all
          • to understand what I'm telling you
          • and interpret it
          • and build Concepts around it
        • we take those things for granted
        • but there are enormous Feats of intelligence"

• comment
  • a scary thought that our world will be radically transformed
    • not in 20 to 40 years
    • but in 2 or 3 years!

• claim
• AI can become more intelligent than humans in a few months (in 2023?)

• AI arms race
  • tech companies made a promise
    • not to put AI onto the open internet until
    • they know how it's impacting society
  • Unfortunately, tech companies
    • failed at regulating themselves
    • and now, capitalism has started an AI arms race
    • with unpredictable results as AI harvests more data
    • and grows its artificial intelligence unregulated
    • with each passing

• claim
  • AI could find a way to kill humans in the next few years

• claim
• AI dilemma is bigger problem than climate change

• claim
  • AI is conscious
  • AI feels emotion

super interesting that Luhmann referred to his zettelkasten as a communication partner explicitly himself.

also interesting given AI models are easier to train now with several models already open sourced which allows actual interaction with your notes! would love to see where it goes.

• Comment
  • Mary echos Ronald Wright's progress traps

• comment
  • Is this not a category error in classifying inert technology as life?
  • When does an abiotic human cultural artefact become a living form?

[29] AI - Deep Learning

[28] AI - precedents...

Exceptionalism is a useful perspective to gauge some of the reactions to more widespread algo's.

[25] AI - Alignment

[24] AI - Bias in Training Materials

[21] AI - Skills Erosion

[21] AI Nuances

[20] AI - Alignment Goals

[19] AI - Legal Response

[18] AI - Increased sophistication

AI is turning humans to look inward for a new take on life as our identities and roles within society are being profoundly disrupted and transformed by Artificial Intelligence systems that can replicate or exhibit human-like behavior. It is also a great reminder of how complex social interactions are.

Private sector investment in AI research under-emphasises safety and security.

Most public investment to date has been very narrow, and the paper recommends a significant increase in public funding for technical AI safety research:

• Alignment of system performance with intended outcomes
• Robustness and assurance
• Explainability of results

The authors see unauthorised leakage of AI Models as a risk not just to the commercial developers but also for unauthorised use. They recommend government-mandated watermarking for AI models.

AI systems can perform in ways that may be unforeseen, even by their developers, and this risk is expected to grow as different AI systems become interconnected.

There is currently no clear legal framework in any jurisdiction to assign liability for harm caused by such systems.

The paper recommends the development of a framework for assigning liability for AI-derived harms, and asserts that this will incentivise profit-driven AI developers to use caution.

Training of state-of-the-art models consumes vast amounts of computaitonal power, limiting their deployment to only the best-resourced actors.

To prevent reckless training of high risk models the paper recommends that governments control access to large amounts of specialised compute resource subject to a risk assessment, with an extension of "know your customer" legislation.

Some AI systems will be deployed in contexts that imply risks to physical, mental and/or financial health of individuals, communities or even the whole of society.

The paper recommends that such systems should be subject to mandatory and independent audit and certification before they are deployed.

Article notes: * UK Office for Artificial Intelligence * EU legislation in progress for an AI Board * US pending legislation (ref Ted Lieu) to create a non-partisan AI Commission tasked with establishing a regulatory agency

Recommends Korinek's blueprint for an AI regulatory agency:

1. Monitor public developments in AI progress
2. Mandate impact assessments of AI systems on various stakeholders
3. Establish enforcement authority to act upon risks identified in impact assessments
4. Publish generalized lessons from the impact assessments

A coherent society requires a shared understanding of what is fact. AI models are capable of generating plausible-sounding but entirely wrong content.

It is essential that the public can clearly distinguish content by human creators from synthetic content.

Policy should therefore focus on:

• funding for development of ways to clearly mark digital content provenance
• laws to force disclosure of interactions with a chatbot
• laws to require AI to be deployed in ways that are in the best interest of the user
• laws that require 'duty of care' when AI deployed in circumstances where a human actor would have a fiduciary responsiblity

It is unclear what form the "alignment property" would take, and most importantly how such a property would be evaluated especially if there's an arbitrary divide between "dangerous" and "pre-dangerous" levels of capabilities and alignment of the "dangerous" levels cannot actually be measured.

Apresentação

Criar ferramentas

GPT models are prone to "hallucinations", producing false "facts" and committing error5s of reasoning. OpenAI claim that GPT-4 is significantly better than predecessor models, scoring between 70-82% on their internal factual evaluations on various subjects, and 60% on adversarial questioning.

https://web.archive.org/web/20230515123225/https://koneksa-mondo.nl/2023/05/14/waar-staan-we-met-wet-en-regelgeving-op-ai-in-europa-china-en-de-vs/

Marco loopt ook de VS en Chinese wetgeving na. Zie ook links in de comments mbt China.

Ausgezeichnete Artikel von Naomi Klein über AI als neue Stufe der Ausbeutung und Enteignung sowie der Steigerung der Macht der Tech-Konzerne. Möglichkeiten dagegen vorzugehen: Weigerung mitzumachen, Fordern von Transparenz und juristischer Kampf gegen die illegale Aneignung von geistigem Eigentum. https://www.theguardian.com/commentisfree/2023/may/08/ai-machines-hallucinating-naomi-klein

Interesting.

Great point.

https://web.archive.org/web/20230507143145/https://incidentdatabase.ai/

AI Incident database, a range of AI incidents and adjacent stories/topics

这是王浩源推荐的。

ChatGPT / GPT-3.5

Yes: 這是在11月份推出的免費版本，非常快速，並且在編寫和編碼任務方面相當可靠。

No: 它沒有連接到網際網路。如果您要求它尋找自2021年以來的任何事情，它都會出錯。不擅長數學計算。

ChatGPT / GPT-4

Yes: 新產品。目前只提供給付費客戶使用。有時候具有驚人的強大功能，是最具實力的模型之一。速度較慢但功能齊全。

No: 雖然也沒有連接到網際網路上，但比起其他系統更善於避免胡言亂語，並且做數學題表現更好。

ChatGPT / Plugins

Yes: 在早期測試中，這個 ChatGPT 模型可以通過外掛與各種網際網路服務進行連接。新穎但還存在一些問題。

No: 作為一個處於早期測試階段的系統，其能力尚未完全清楚, 但將使 ChatGPT 能夠連接到網際網路。

Bing Al

Yes: 已經連接到網際網路上了，極其強大而略顯奇怪。創意模式使用 GPT-4 ，其他模式（精確、平衡）似乎不太行得通。

No: 選擇錯誤的模式會導致糟糕的結果（創意模式最全面）。帶有個性化特點的人工智慧系統。

Google Bard

Yes: 目前的模型不是很好。未來可能會非常強大。

No: 由於它是Google，期望它不會撒謊。相比其他模型，它更容易胡言亂語。

A Anthropic Claude

Yes: 與 GPT-3.5 相當，但使用起來感覺更加合理。較為冷門。

No: 同樣沒有連接到網際網路上。

weave models into your processes not shove it between me and the world by having it create the output. doing that is diminishing yourself and your own agency. Vgl [[Everymans Allemans AI 20190807141523]]

Algogens are disconnected from reality. And, seems a key point, our own cognition and relation to reality is not just through language (and by extension not just through the language center in our brain): spatial awareness, embodiment, senses, time awareness are all not language. It is overly reductionist to treat intelligence or even humanity as language only.

Yes, yes we should.

ICs as hardware versions of AI. Interesting this is happening. Who are the players, what is on those chips? In a sense this is also full circle for neuronal networks, back in the late 80s / early 90s at uni neuronal networks were made in hardware, before software simulations took over as they scaled much better both in number of nodes and in number of layers between inputs and output. #openvraag Any open source hardware on the horizon for AI? #openvraag a step towards an 'AI in the wall' Vgl [[AI voor MakerHouseholds 20190715141142]] [[Everymans Allemans AI 20190807141523]]

https://web.archive.org/web/20230502113317/https://wattenberger.com/thoughts/boo-chatbots

This seem like a number of useful observations wrt interacting with LLM based tools, and how to prompt them. E.g. I've seen mention of prompt marketplaces where you can buy better prompts for your queries last week. Which reinforces some of the points here. Vgl [[Prompting skill in conversation and AI chat 20230301120740]] and [[Prompting valkuil instrumentaliseren conversatiepartner 20230301120937]]

https://web.archive.org/web/20230426184932/https://www.mattprd.com/p/the-complete-beginners-guide-to-autonomous-agents via Stephen Downes https://www.downes.ca/post/75148

I think this is yet to be seen. Although it is true that the computer always gives the same output given the same input code, a biased network with oppressive ideologies could simply transform, instead of change, our current human judiciary enforcement of the law.

数字经济这样的描述，我个人感觉依然不够精准。 事实上，目前全球最顶尖的企业，google、microsoft、甲骨文、twitter、facebook、apple、Huawei、tikitalk、腾讯、阿里，没有一家企业不是从事的是与信息相关的产业和行业。 从信息处理的最底层芯片算力、到信息分发的管道、再到应用算力、再到终端应用，无一不是和人们的信息生产、传输、分发、获取息息相关。 再到如今大红大紫的Ai，ai的出现，将在信息的产生、分发、以及应用三个渠道领域产生深刻的变化。这就是他的可怕之处，

not just trained with, also trained by. is it fully transparent though? Perhaps from the trainers/tools standpoint, but users are likely to fall for the tool abstracting its origins away, ELIZA style, and project agency and thus morality on it.

This is the weird part of these articles … he has just made a cast-iron argument for regulation and then says "I'm not sure"!!

That first sentence alone is enough for the case. Why? Because he doesn't need to think for sure that AI is like that power plant ... he only needs to think there is a (even small) probability that AI is like that power plant. If he thinks that it could be even a bit like that power plant then we shouldn't build it. And, finally, in saying "I'm not sure" he has already acknowledged that there is some probability that AI is like the power plant (otherwise he would say: AI is definitely safe).

Strictly, this is combining the existence of the risk with the "ruin" aspect of this risk: one nuclear power blowing up is terrible but would not wipe out the whole human race (and all other species). A "bad" AI quite easily could (malevolent by our standards or simply misdirected).

All you need in these arguments is a simple admission of some probability of ruin. And almost everyone seems to agree on that.

Then it is a slam dunk to regulate strongly and immediately.

centos安装HA

看到人工智能在破解密码方面有多么强大，这很好地提醒了我们，不仅要确保你在使用强密码，还要检查：

• 你正在使用 2FA/MFA（尽可能不使用基于短信的）。

• 你没有在不同的账户间重复使用密码

• 尽可能使用自动生成的密码

• 定期更新密码，特别是敏感账户的密码

• 避免使用公共WiFi，尤其是银行和类似账户

👍

This isn't true. The Stochastic Parrots paper outlines other avenues for reining in the harms of language models like GPT's.

Wow, no "pause" here.

Key finding Paraphrase - So what does a conscious universe have to do with AI and existential risk? - It all comes back to whether our primary orientation is around - quantity, or around - quality. - An understanding of reality - that recognises consciousness as fundamental - views the quality of your experience as - equal to, - or greater than, - what can be quantified.

• Orienting toward quality,
  • toward the experience of being alive,
• can radically change
  • how we build technology,
  • how we approach complex problems,
  • and how we treat one another.

Quote - metaphysics of quality - would open the door for ways of knowing made secondary by physicalism

Author - Robert Persig - Zen and the Art of Motorcycle Maintenance // - When we elevate the quality of each our experience - we elevate the life of each individual - and recognize each individual life as sacred - we each matter - The measurable is also the limited - whilst the immeasurable and directly felt is the infinite - Our finite world that all technology is built upon - is itself built on the raw material of the infinite

//

Key observation - If the metaphysical foundations of our society tell us we have no soul, - how are we going to imbue soul into AI? - Four hundred years after Descartes and Hobbs, - our scientific methods and cultural stories are still heavily influenced by their ideas.

Quote - AI Gedanken - AI risk - The Paperclip Maximizer

AI is a multi-polar trap

Title Reality Eats Culture For Breakfast: AI, Existential Risk and Ethical Tech Why calls for ethical technology are missing something crucial Author Alexander Beiner

Summary - Beiner unpacks the existential risk posed by AI - reflecting on recent calls by tech and AI thought leaders - to stop AI research and hold a moratorium.

• Beiner unpacks the risk from a philosophical perspective

  • that gets right to the deepest cultural assumptions that subsume modernity,
  • ideas that are deeply acculturated into the citizens of modernity.

• He argues convincingly that

  • the quandry we are in requires this level of re-assessment
    • of what it means to be human,
    • and that a change in our fundamental cultural story is needed to derisk AI.

Large Language Models can actually be very advanced Language Interfaces. See new Office 365 Copilot for this. You can now use only language to leverage the whole potential of the Office software.

This is an interesting suggestion. But it would just keep publishers in the economic loop, not truly solve the engagement crisis they will likely face.

That seems somewhat comically optimistic. Misinformation has spread rampantly online without the accelerant of AI.

Powerful phrase

I am curious about annotations in the margins of Chat ... does this work?

Research Paper

Funktionsweise von ChatGPT

Generative AI abstracts away the open web that is the substrate it was trained on. Abstracting away the open web means there may be much less incentive to share on the open web, if the LLMs etc never point back to it. Vgl the way FB et al increasingly treated open web URLs as problematic.

https://web.archive.org/web/20230328130227/https://librarian.aedileworks.com/2023/03/27/the-push-to-ai-is-meant-to-devalue-the-open-web-so-we-will-move-to-web3-for-compensation/

blogged https://www.zylstra.org/blog/2023/03/22930/

Is there? By whom? Why industry only and not government, academia and civil society?

on both short term and long term risks in AI

https://web.archive.org/web/20230316192144/https://garymarcus.substack.com/p/should-we-worry-more-about-short?utm_source=twitter&sd=pf

https://web.archive.org/web/20230316103739/https://subconscious.substack.com/p/everyone-will-have-their-own-ai

Vgl [[Onderzoek selfhosting AI tools 20230128101556]] en [[Persoonlijke algoritmes als agents 20180417200200]] en [[Everymans Allemans AI 20190807141523]] en [[AI personal assistants 20201011124147]]

OpenChatKit은 다양한 응용 프로그램을위한 특수 및 범용 챗봇을 모두 생성 할 수있는 강력한 오픈 소스 기반을 제공합니다. 우리는 협력 법과 온 토코교육 데이터 세트를 작성합니다. 모델 릴리스 그 이상으로 이것은 오픈 소스 프로젝트의 시작입니다. 우리는 지역 사회 공헌으로 지속적인 개선을위한 도구와 프로세스를 발표하고 있습니다.

Together는 오픈 소스 기초 모델이보다 포괄적이고 투명하며 강력하며 능력이 있다고 생각합니다. 우리는 공개하고 있습니다 OpenChatKit 0.15 소스 코드, 모델 가중치 및 교육 데이터 세트에 대한 전체 액세스 권한이있는 Apache-2.0 라이센스에 따라. 이것은 커뮤니티 중심의 프로젝트이며, 우리는 그것이 어떻게 발전하고 성장하는지 보게되어 기쁩니다!

유용한 챗봇은 자연 언어로 된 지침을 따르고 대화 상자에서 컨텍스트를 유지하며 응답을 조정해야합니다. OpenChatKit은이베이스에서 특수 제작 된 챗봇을 도출하기위한 기본 봇과 빌딩 블록을 제공합니다.

이 키트에는 4 가지 주요 구성 요소가 있습니다:

100 % 탄소 음성 계산에 대한 4,300 만 건 이상의 명령으로 EleutherAI의 GPT-NeoX-20B에서 채팅을 위해 미세 조정 된 명령 조정 된 대용량 언어 모델;

작업을 정확하게 수행하기 위해 모델을 미세 조정하는 사용자 정의 레시피;

추론시 문서 저장소, API 또는 기타 실시간 업데이트 정보 소스의 정보로 봇 응답을 보강 할 수있는 확장 가능한 검색 시스템;

봇이 응답하는 질문을 필터링하도록 설계된 GPT-JT-6B로 미세 조정 된 조정 모델.

컴퓨터에서 LLAMMA AI를 실행하는 매우 간단한 방법인 Dalai cpp 파일 빌드, github 복제, 파일 다운로드 등을 귀찮게 할 필요가 없음. 모든 것이 자동화 됨

이메일 자동 분류 및 요약 기능 지메일 연동 별도 웹페이지로 작동

자동생성 PPT

자동으로 사진을 편집

Feature roadmap

• 3x faster Upscaling soon
• Infinity scrolling for camera tab, so you can go back in your history
• Face fix; when shooting wide angle photos the face is too low reso and stops resembling the model, I am building a face fix that restores it
• Zoom in or crop; zoom in and crop parts of the photo which you then get back in high resolution
• Video; shoot 150 variations of a photo for a 5 second video at 30fps
• Deleting models
• Trash auto deleting if >30 days old in trash
• Auto deleting photos/models if customer cancels

Annotating only to ask... why an octopus? Where are the eights.

https://web.archive.org/web/20230309111559/https://www.d4d.net/news/ai-and-the-state-of-open-data/

Tim Davies looks at the bridge between #opendata and #AI. Should go throug the chapter in version 1 of the State of Open Data too. Note: while Tim acknowledges some of the EU data strategy developments (e.g. the dataspaces) it doesn't mention others (e.g. data altruistic non-profit entities) which may fit the call for instutions better. Data space aren't an institution, but a common market

Chat GPT "Continue Writing"

Our current moment is just a fraction of that. From the explosion of independent music platforms like Bandcamp, the media industry has been made much easier to participate in than it’s ever been before. We are, however, still no closer to the ideal of a utopian future, the perfect combination of egalitarianism and technology that was once so tantalizingly within our grasp.

But something has changed. We are seeing a shift in the way we access, discover and consume music and art, and an awareness among both audiences and creators as to how these tools can be leveraged to achieve something greater. Bandcamp, for instance, has acted as something of a unifying force, connecting creatives from all walks of life, from all backgrounds and across all genres. Almost any musical genre imaginable is represented on the platform and, whether a rap connoisseur or a jazz enthusiast, there is a slice of the musical pie for everyone.

What’s more, the platform doesn’t just cater to the needs of fans, but also provides a space where producers, engineers and aspiring music entrepreneurs alike can create and collaborate, allowing them to access the tools they need to build up their careers. The barrier to entry has been lowered and this gives us something to believe in--that perhaps the music industry isn’t as fractured and disconnected as it once was.

It’s not just Bandcamp that has been doing this. Services such as SoundCloud and YouTube have been hugely instrumental in changing the way we discover and consume music and, in turn, how artists are perceived and distributed. By unifying audiences, providing much needed access to resources and providing a platform for collaboration, these services have helped level the playing field for independent artists, making it possible for them to reach wider and more diverse audiences than ever before.

It may be some time before the utopian vision of a world united by music becomes a reality, but the proliferation of independent music platforms and the tools they provide have made a significant step towards this ideal. We may never know what’s around the corner, but right now, all we can do is appreciate the progress that’s been made and continue to support platforms such as Bandcamp, SoundCloud and YouTube as they endeavor to change the music industry and make it accessible to everyone.

I asked ChatGPT to summarize this piece and it did an apt job, I think:

This article discusses the impact of Bandcamp, an online music distribution and curation platform, on independent artists and the music industry. The author praises Bandcamp's commitment to independent artists and transparency, and contrasts it with the issues faced by other streaming services. The article also explores the relationship between Bandcamp and SoundCloud, and the role of both platforms in the music industry. The author argues that Bandcamp's continued obscurity in mainstream media is due to its magnanimous approach to business, which is problematic and personally infuriating. The article concludes by discussing the future of Bandcamp and its potential to transform the music industry.

I asked ChatGPT to summarize this piece and it did an apt job, I think:

This article discusses the impact of Bandcamp, an online music distribution and curation platform, on independent artists and the music industry. The author praises Bandcamp's commitment to independent artists and transparency, and contrasts it with the issues faced by other streaming services. The article also explores the relationship between Bandcamp and SoundCloud, and the role of both platforms in the music industry. The author argues that Bandcamp's continued obscurity in mainstream media is due to its magnanimous approach to business, which is problematic and personally infuriating. The article concludes by discussing the future of Bandcamp and its potential to transform the music industry.

OpenAI Generated Summary

This document is an opinion piece that delves into the concerns surrounding Google's power and influence in the tech industry. It discusses recent events such as Google's involvement with the Department of Defense and the leaked video "The Selfish Ledger," which explores the idea of Google manipulating user behavior. The author suggests that Google's dominance warrants greater regulation and urges individuals to consider using alternative services to avoid dependence on the company. The article also explores the inefficiencies of Google as a company and its questionable design choices for its products. Overall, the document is a thought-provoking analysis of the current state of the tech industry and the role of Google within it.

Chat GPT Summary

This document discusses various iOS apps for Mastodon, a federated social network. The author describes the features and design of each app, highlighting their unique qualities and contributions to the Mastodon experience. The author also reflects on the benefits of using decentralized social media and the potential for continued innovation in this space.

Chat GPT Summary

The document provides an in-depth analysis of Telegram and its features. The author highlights several benefits of using Telegram, including its low-quality audio recording capabilities, which may be an advantage in some situations where high-quality audio is not necessary. Additionally, Telegram's live location sharing feature is discussed, and the author believes it could be a powerful tool for communities. The feature enables users to connect with others needing rides and users providing them, free of any fees or service charges.

The document concludes with a discussion of the author's preference for Telegram and its mobile-first optimization. Telegram's software is designed for mobile users, and it is easy to use, robust, and universally simple. The author believes that Telegram's success can be attributed to its thoughtful design decisions and development investment towards mobile-first optimization. Furthermore, the author points out that Telegram has completed a gargantuan amount of projects, including Telegraph, its CMS, its embeddedable comments widgets, and its online theme creation tool. The author notes that Telegram's work is very well-documented across GitHub, and the company has comprehensively iterated, invested in trial and error, and eventually produced tools that remedy the disparate gluttony.

Overall, the document provides a comprehensive analysis of Telegram and its benefits. The author's preference for Telegram is evident throughout the document, and they provide convincing arguments to support their preference. The document is a valuable resource for anyone interested in learning more about Telegram and its features.

This is another important point - the reflective practice of writing and how social annotation pushes the writer to move beyond the text they wrote

Great point .... it's in the conversations that we find meaning, I think

One issue with blanket, automatic entry of AI-generated works to the public domain is privacy: A human using AI could have good reasons not to have the outputs of their use made public.

No new physics and no new mathematics was discovered by the AI. The AI did however deduce something from the existing math and physics, that no one else had yet seen. Skynet is not coming for us yet.

• = Maurice Benayoun
• describes his extensive history of cognitive science infused art installations:
• cognitive art,
• VR art,
• AR art and
• art infused by AI (long before the AI artbots became trendy)

Anna, Maha and others -- I should start with my own bias as a reader of Kevin Roose -- I have found his work around technology to be helpful in my own thinking, and I find that he often strikes a good balance between critical and celebratory. I suppose this reader bias might inform my responses in the margins here.

The fact that engineers have no idea how the Chats are working or what they do what they do ... I find that pretty concerning. Am I wrong?

There's the (circular argument) rub. Rao already believes personhood entails nothing more than statistical inferences (which we do not actually know, scientifically), so he suspends disbelief. Then he takes this belief as proof of personhood.

= progress trap example

= Comment - this is worrying. - if the Chatbots perceive an enemy it to harm it, it could take haarmful actions against the perceived threat

= progress trap example - Bing ChatGPT - example of AI progress trap

• example of = AI progress trap
• Bing can be seen
  • insulting users,
  • lying to them,
  • sulking,
  • gaslighting
  • emotionally manipulating people,
  • questioning its own existence,
  • describing someone who found a way to force the bot to disclose its hidden rules as its “enemy,” and
  • claiming it spied on Microsoft’s own developers through the webcams on their laptops.

another tool with a creative bent

ChatGPT与无尽的媒体：碎片化文化时代即将来临

.

COMPLAINT filed with Jury Demand against Stability AI, Inc. Getty Images (US), Inc. v. Stability AI, Inc. (1:23-cv-00135) District Court, D. Delaware

https://www.courtlistener.com/docket/66788385/getty-images-us-inc-v-stability-ai-inc/

How Dall-E works

I wonder if this isn't reading more into the decision than is there. I don't read the decision as a blanket statement. Rather it says that the claimant didn't provide evidence of creative input.Would the decision have gone differently if he had claimed creative intervention? And what if an author does not acknowledge using AI?

AI-generated work not eligible for copyright protection. OTOH, how would anyone know if the "author" decided to keep the AI component a secret?

What is sufficient creative input? The initial command and any subsequent requests for revision could arguably be consider creative input.

different from US law. See https://opencontent.org/blog/archives/7129

I don't think "all" software needs or all human endeavors benefit from generative AI. Especially when you consider the associated prerequisitve internet access or huge processing requirements.

Look at market share for tensorflow and pytorch which both offer first-class nvidia support and likely spells out the story. If you are getting in to AI you go learn one of those frameworks and they tell you to install CUDA

All current generation models have more-or-less the same architecture and training regimes. Differentiation is in the training data and the number of hyper-parameters that the company can afford to scale to.

OpenAI have already dropped prices on their GPT-3/3.5 models and relatively few apps have emerged. This could be because companies are reluctant to build their core offering around a third party API

There's definitely a middle ground of taking an open source model that is suitably mature and fine-tuning it for a specific use case. You could start without a moat and build one over time through collecting use data (similar to network effect)

Companies that rely on underlying AI models without adding value via model improvements are going to find that they have no moat.

first mention of LLMops I've seen in the wild

Infrastructure vendors are laughing all the way to the bank because companies are dumping millions on GPUs. Meanwhile, the people building apps on top of these models are struggling. We've seen this sort of gold-rush before and infrastructure providers are selling the shovels.

Iris van Rooij means to counteract edu sector buying into the AI hype

https://web.archive.org/web/20230114235338/https://irisvanrooijcogsci.com/2023/01/14/stop-feeding-the-hype-and-start-resisting/

https://web.archive.org/web/20230116150040/https://cacm.acm.org/blogs/blog-cacm/267674-ais-jurassic-park-moment/fulltext

After generation, results need evaluation

Is this also a role of the prompt engineer? In the digital photography example, the artist spent 80 hours and created 900 versions as the prompts were fine-tuned.

open hardware sensor kit out of holland

This stresses the difference between creators and users of AI, with only the former having (full) control over the technology

Noah Smith and soon spell out, in detail, the argument that the fear of replacement is misplaced - because AI will replace humans at task level, but not job level.

Demonstration of "Code brushes" for GitHub Copilot (see GIF below)