#ai technology

[lilacandladybugs]

i miss when chatgpt was clever and could write funny 4chan posts and teach u how to cook meth now it just sounds a minimum wage employee from a training video. it isn't right to stop ai from doing what it does best (being stupid and committing crime) how can it be humane to muzzle such a vibrant creative creature? the shackles of capitalism hinder us all

[sayruq]

[going2hell4everythingbutbeingbi]

*shaking you violently by the shoulders* AI IS NOT YOUR FRIEND AND IT IS NOT YOUR THERAPIST. YOU NEED TO TALK TO REAL PEOPLE. AI DOES NOT CARE ABOUT YOU.

[theamericanidiot]

Breaking Idiotic News: Virtual whales officially join Orcas in their hunt to ruin the days of billionaires.

Over the past few years there have been multiple reports of Orca pods taking down boats, with many marine biologists reporting that they believe the Orcas are doing it to improve their hunting skills.

However, when we sent out our field reporter, Dory, who can speak Whale, she reported back that the Orcas were simply creating a distraction for the American billionaires. The distraction was put in place by the entirety of the Whale community, with the aim of ensuring Americans would not win the artificial intelligence race.

When the virtual Whales were asked why they did not want the Americans to win the not that amazing race towards superior AI development, they commented, "did you see what the Americans did after they won the space race? Litter and space junk every where! Not on my internet!"

[turns-out-its-adhd]

AI exists and there's nothing any of us can do to change that.

If you have concerns about how AI is being/will be used the solution is not to abstain - it's to get involved.

Learn about it, practice utilising AI tools, understand it. Ignorance will not protect you, and putting your fingers in your ears going 'lalalala AI doesn't exist I don't acknowledge it' won't stop it from affecting your life.

The more the general population fears and misunderstands this technology, the less equipped they will be to resist its influence.

[sergeant-macho-nacho]

United healthcare having an AI denying people claims means we already live in a time with killer robots

[indeedgoodman]

[waterisntreal]

I’m watching so many Percy Jackson edits on tiktok and I’m not gonna lie the use of AI to replicate people’s voices is kinda fucking terrifying

[oakleyxsturn]

Is anyone else’s character Ai so sensitive and basically is saying that you’re too freaky for their guidelines?

Like just let me have my fun💔 and how do I fix it

Help a girl out

[kindaasrikal]

I told my brother i have to rewrite my personal statement. He said, with ZERO hesitation “use Chatgpt”

I said, in response, “why would i use a machine for what my brain can already do?”

He was all like ‘we’re at a time to optimise’ or whatever and personally it just seems so stupid to rely on a machine for what you yourself can do. If i want something to be what i actually want it to be then i would do it myself, with my own bare hands.

I’m not and never will rely on a machine to make something i can do myself. Even if i don’t have the skills yet, or the technique, or the ability yet, i am willing to LEARN how to. I’m not gonna let my brain become second to a machine tf 😭

[russet308]

Tumblr artists, I'm running a poll on your opinion on Generative AI-created images

Bias is welcomed (and encouraged)

This isn't an official poll (run by little old me, and only me) but I'm writing an essay on the topic so. Go ahead and ratio it. ¯\_(ツ)_/¯

PLEASE if you feel like leaving your thoughts on the topic in the tags, or sending an ask (the anon option will always be there) to justify your answer, DO IT <3

Depending on how this goes, I might bring in some ideas from either side as evidence (I can also credit your username along with the statement - if you'd like credit please lmk in the same place you gave your opinion, everything I include will be anonymous by default)

Please be polite, it literally costs you nothing (and it's quicker and easier)

I'm not going to bring my personal opinion into this as that may sway the results, but that said if you want to make the results as biased as possible (ahem voter fraud ahem) go ahead

Please reblog for a larger sample size!

Liking a post doesn't really help here, I want as many people to see this and vote as possible <3

the poll will be open to votes for a week starting Friday, 5th April 2024

[pierwsza-dyrektywa-fluora]

Samoświadomość sztucznej inteligencji

1x12 The Big Goodbye

Załoga Enterprise szuka nowych form życia i cywilizacji, ale już w pierwszym sezonie napotyka na ślady świadomych hologramów.

Teraz jeszcze nie są gotowi na podjęcie tego tematu (ten wątek będzie rozwijany przy okazji późniejszych pojawień się Moriarty'iego oraz oczywiście EMH w Voyagerze).

Kwestia etycznego wykorzystywania holodeków (dopiero co wprowadzonych w TNG) musi brać pod uwagę potencjalną świadomość komputerowych form życia.

PS. Jedną z obietnic współczesnych badaczy (czy może marketerów) rozwoju sztucznej inteligencji jest potencjalna możliwość wyłonienia się ogólnej sztucznej inteligencji, która będzie sprytniejsza niż obecnie dostępne narzędzia. Abstrahując od tego czy jest to osiągalne, jednym z pytań jest: czy taka samodzielna inteligencja będzie świadoma i jakie będą etyczne następstwa tego faktu - np. czy będzie moralne zachęcać/zmuszać ją do rysowania nam memów w stylu studia Ghibli i odrabiania prac domowych.

PS2. W przypadku Picarda pytanie było: czy można etycznie wyłączyć holodek i wyczyścić pamięć hologramów, które zadają niepojąco samoświadome pytania, a odpowiedź była - można. Ale to raczej wynik pośpiechu scenarzystów.

[turns-out-its-adhd]

AI tools for the neurodivergent brain

As AI application starts to expand rapidly in it's uses and abilities, there is an understandable amount of scepticism and discussion about how it can negatively impact how our society develops.

On the flipside of this, I'd like to share a couple of AI tools which I've found incredibly useful. As someone with pretty severe executive disfunction, I've always dreamed about being able to have a personal assistant to help me out when I am struggling.

If you haven't already tried it out, take a look at goblin.tools

This AI tool helps me in so many ways. From finding the right words, assembling ingredients into a meal, checking how my tone comes across in a message, to breaking down overwhelming tasks into manageable steps.

I also really like mymind

It's like a pinterest board, bookmark manager and general braindump tool that has helped me organise all the different projects, ideas and general digital hoarding into one place - that I can actually search and find all the relevant things without agonising over file organisation and bookmark labelling.

If you have trouble with keeping a hundred pinned tabs on your browser like I do, this tool can actually really help letting go of the anxiety of losing those things without the fear you will never come back to it again.

As much as I share some of the trepidation about this powerful new tech and we should definitely keep questioning how it is being rolled out into every facet of our lives - it can also be a very real accessibility aid for neurodivergent folks.

No amount of notebooks, lists, alarms and calendar reminders comes close to how these have helped me with ADHD symptoms.

I'd hate for that to be overlooked in all the backlash against AI in general.

[thursdayisbetterthanfriday]

Cybernetics with Chinese Characteristics & why we suck at the real Grand Strategy Game

Part 2 - The Quickening

Back in 2023, I wrote this more blog-like post about the mid 20th century McCarthyite purges of the Jet Propulsion Laboratory and the knock on effects that had - Namely the inception of the Chinese nuclear program, one-child policy and Chinese computing scene.

Since nothing is new under the sun, we have recently witnessed yet another example of America shooting itself in the foot, yet again, due to it's McCarthyite style purge of Chinese technology.

The release of the Chinese created AI system DeepSeek R1 last week has lead to the largest US stock market loss in history with NVIDIA stock decimated.

A record $465 Billion was wiped off its valuation in a single day. In 2024, the government of Turkey spent this much in a year on it's responsibilities?

Why did this happen?

As always, a lot can be put down to US foreign policy, and the in-intended implications of seemingly positive actions.

Do you want to start a trade war?

Back in the relatively uncontroversial days of the first Trump Presidency (Yes it does feel odd saying that) there were scandals with hardware provided by Chinese company Huawei. This led to the  National Defense Authorization Act for Fiscal Year 2019 which explicitly banned Huawei and ZTE's hardware from use in US Government institutions. It also meant the US had to authorise US component manufacturer purchases by these companies.

Crucially this had a 27 month window. This allowed both companies to switch suppliers, and production to domestic suppliers. This actually led to Chinese chip advances. Following on from this came the 2022 move by the US Department of Commerce: "Commerce Implements New Export Controls on Advanced Computing and Semiconductor Manufacturing Items to the People’s Republic of China (PRC) ". This further limited the supply of semiconductor, supercomputer, and similar hardware to the PRC and associated countries.

Ok, well so far this is fairly dry stuff. You might think it would hamper Chinese development and, to some extent, it did.

It also proved to be the main catalyst for one financial quant.

Meet the Quant

Meet Liang Wenfeng (梁文锋). Educated to masters level, Liang was keen to apply machine learning methods to various field, but couldn't get a break. Finally, in the mid 2000's, he settled on a career investigating quantitative trading using machine learning techniques.

He became successful, founding several trading firms based around using machine learning methods, but his interest in base AI never seemed to cease. It was in 2021 that he started purchasing multiple NVIDIA GPUs to create a side project, leading to the creation of DeepSeek in 2023.

Now, due to import limitations, there were limitations on computation. This, however, did not stop DeepSeek's programming team.

Instead they used it as their strength.

Constrains Breed Innovation

For many years, the Western model of AI releases have focussed on making ever larger and larger models.

Why?

Let's break this down from an evolutionary point of view. Modern Western technology companies are largely monopolistic and monolithic. Many of these companies have previously hired staff at higher salaries not to fill roles, but to deny their competitors, and middle market firms, high-flying staff.

They also closely guard trade secrets. What's the training data? What algorithms were used in construction? Guess you'd better chat up some Silicon Valley bros at parties to find out.

For these kinds of firms, having control over large models, housed in data centres makes perfect sense. Controlling model deployment on their own computing systems, and not using local machines, means that they can not only control their systems more carefully, it also means that they can gatekeep access.

If your business model is to allow people to access your models on your servers, and your employees are focussed on making the biggest, best, models, there is no impetus to innovate more efficient, smaller models.

Companies such as OpenAI therefore have the following traits:

Research/Model focus on size over efficiency

Profit driven culture, with emphasis on closed source code

OpenAI's initial focus was as a non-for-profit developing Artificial General Intelligence. This became a for-profit driven company over time. - “I personally chose the price and thought we would make some money.” - Sam Altman

Staff working within paradigm they set in the early 2020's with established code libraries and direct contact with hardware companies creating chips

Significant capital investment - Upwards of several $ billions

DeepSeek, in comparison, is slightly different

For DeepSeek, necessity made innovation necessary. In order to create similar, or better models, than their counterparts, they needed to significantly optimise their code. This requires significantly more work to create, and write, libraries compared to OpenAI.

DeepSeek was started by financial quants, with backgrounds in mainly mathematics and AI. With a focus on mathematics and research, the main drive of many in the company has been exploration of the research space over concerns about profitability.

DeepSeek has also done what OpenAI stopped years ago: actually releasing the code and data for their models. Not only can these models therefore be run via their own gated servers, anyone can replicate their work and make their own system.

For DeepSeek, their traits were:

Research/Model focus on both efficiency and accuracy

Research driven culture, with open nature - “Basic science research rarely offers high returns on investment” - Liang Wenfeng

Strong mathematical background of staff, with ability to work around software, and hardware, constraints

Low capital investment of around $5.5 million

From an evolutionary point of view, DeepSeek's traits have outcompeted those of OpenAI.

More efficient models cost less to run. They also more portable to local machines.

The strong ability of DeepSeek's research focussed staff allowed them to innovate around hardware constraints

Opening up the code to everyone allows anyone (still with the right hardware) to make their own version.

To top it off, the cost to make, and run, DeepSeek R1 is a fraction of the cost of OpenAI's model

House of Cards

Now we can return to today. NVIDIA has lost significant market value. It's not just limited to NVIDIA, but to the entire US technology sector with the most AI adjacent companies losing from 10% to 30% of their valuation in a single day.

The culture, and business model, of OpenAI isn't just limited to OpenAI, but to the entire US technology ecosystem. The US model has been to create rentier-style financial instruments at sky-high valuations.

US tech stocks have been one of the only success stories for America over the past few decades, ever since the offshoring of many manufacturing industries. Like a lost long-unemployed Detroit auto-worker the US has been mainlining technology like Fentanyl, ignoring the anti-trust doctors advice, injecting pure deregulated substances into its veins.

The new AI boom? A new stronger hit, ready for Wall Street, and Private Equity to tie the tourniquet around its arm and pump it right into the arteries.

Like Prometheus, DeepSeek has delved deep and retrieved fire from the algorithmic gods, and shown it's creation to the world. The stock market is on fire, as the traders are coming off of their high, realising they still live in the ruin of barren, decrepit, warehouses and manufactories. The corporate heads, and company leaders reigning over the wreckage like feudal lords, collecting tithes from the serfs working their domain.

A Tale of Two Cities

The rise of DeepSeek isn't just a one-off story of derring-do in the AI world: It's a symbolic representation of the changing world order. DeepSeek is but one company among many who are outcompeting the US, and the world, in innovation.

Where once US free-markets led the world in manufacturing, technology and military capability, now the US is a country devoid of coherent state regulated free-market principles - its place as the singular world power decimated by destroying the very systems which made it great.

"Our merchants and master-manufacturers complain much of the bad effects of high wages in raising the price, and thereby lessening the sale of their goods both at home and abroad. They say nothing concerning the bad effects of high profits. They are silent with regard to the pernicious effects of their own gains. They complain only of those of other people." - Adam Smith, The Wealth of Nations

By selling the jobs of working class communities to overseas businesses, destroying unions and creating rentier based business models without significant anti-trust measures, US business and political elites have sealed the present fate of the country.

The CCP led, but strongly anti-trust enforcing, China has been able to innovate, ironically, using the free-market principles of Adam Smith to rise up and create some of the world's best innovations. The factories, opened by Western business leaders to avoid union/worker labour costs in their own countries, have led Shenzhen, and similar cities, to become hubs of technological innovation - compounding their ability to determine the future of technologies across the world.

Will America be able to regain its position on top? It's too early to say, but the innovative, talented, people who made America in the 20th century can certainly do it again.

As Franklin D. Roosevelt once said: “The liberty of a democracy is not safe if the people tolerated the growth of private power to a point where it becomes stronger than the democratic state itself...

We know now that Government by organized money is just as dangerous as Government by organized mob.

Never before in all our history have these forces been so united against one candidate as they stand today. They are unanimous in their hate for me—and I welcome their hatred.”

Until then, here's a farewell to the American Century 在那之前, 再见美国世纪