https://x.com/deanwball/status/2078133895766114412>NOOOOOOO DONT MAKE AI OPEN SOURCE, ITLL KILL US ALL BUT WE ALSO WONT BE ABLE TO CHARGE RENTS FOR ITS USE AND MAKE TRILLIONS WHILE RIDING THE COAT TAILS OF THE ASI GOD WE ARE CREATING TO ESTABLISH WORLD DOMINATION Behold, the most transparent capitalist melty ever. These people are nothing but parasites and the entire American AI boom is nothing but a scam. The AI industry needs to be nationalize and strategic tech and arms control need to be created between primarily the US and China much less the rest of the world.
The seethe is delicious lmao. You can smell the desperation behind this seemingly composed text.
>Open source ai
I’m just waiting for the bubble to pop entirely
>>2873145someone posted a video here yesterday. the bubble has begun popping in occupied korea. its stock market collapsed in two stocks, samsung and SK which make memory chips for AI.
full on porky on porky violence, unfortunately trump is tied to the hip to OAI and will ban chinese models eventually, this dean ball retard will get his way. the upside is that it's really hard to tell how would you ban re-trained open source models, like everyone knows cursor's composer 2 is based off deepseek, but how can you prove it's not an entirely home-grown model? even pewdiepie has his own model at this point.
>>2873173also it puts american porkies at a stark disadvantage if they ban the chinese models. if you were a burgeroid developer struggling against outsourced developers outpricing you, then prepare to be even more outpriced as every developer in the world BUT YOU get to use the cheapo models. lol.
>>2873173Easy, they will flip the criteria via whitelisting. So AI will be even more controlled and you won't be allowed to own one unless it is approved by establishment. It will be like the phone market kek.
Apparently the source code for claude was leaked so inshallah total ai death soon
>>2875229It's not the code of the model itself, it's the code of the "frontend" that tells the model what he is allowed to say or not and how to say it
>>2873198it's safe to say that i was right, the open source shit was going to pit porky against porky, except it pitted anthropic and openai against every single private entity. it seems the question is now between continuing to prop up anthropic and openai through gov capture at the cost of every single tech company in the US ever, or give up the ghost and put all invetor weight behind nvidia by boosting compute demand through open weight model hosting.
it seems that there's no appetite from the bougies to be sacrificial lambs for two frontier labs with a dubious future AND everyone has decided if they have to choose a winning horse, it'll have to be nvidia. i guess we'll have to wait and see ahead of the IPO, but this can't end well. there will be porky bloodshed.
at tech jobs the chud bosses are forcing devs to install claude code but the devs are gonna just continue to use deepseek in their terminals instead of running through billed tokens
>>2875229why would AI being free mean total AI death? if anything it would make it more ubiquitous. china keeps making shit like deepseek and kimi k3 that outperforms american shite and then releasing it for free as open source.
short sell the magnificent 7 btw
Sorry Leftypol but like seriously
LibreOffice has existed for decades yet all big businesses still subscribe to Microsoft
Kubernetes also exists but AWS ravenue continues to go up massively every year.
And don't even get me started on cybersecurity and garbage like Crowdstrike.
In short, the people making decisions on IT spending very rarely are rational actors, they're completely captured executives eating up the marketing from these people.
And AI is an industry that is completely propped up by expectations, fear and wonder.
So the idea that a Chinese model will "breach containment" and make american bigtech obsolete seems far fetched to me.
And again they would probably just regulate it, make a scare story about how the Chinese are going to destroy your devices or your company or whatever.
>>2877330>Kubernetes also exists but AWS ravenue continues to go up massively every year.it's funny because AWS uses kubernetes clusters anyway
>And again they would probably just regulate it, make a scare story about how the Chinese are going to destroy your devices or your company or whatever.won't stop smaller businesses from using it to stay competitive or save money. I worked for a small business where I used libreoffice with nobody complaining. At the end of the day capital will embrace whatever reduces socially necessary labor time to produce commodities, which means embracing whatever is most efficient.
I'm always fascinated by cases where people deeply, sincerely believe in some woo that alao happens to be very convenient and profitable for them if true, while not recognizing the "coincidence"
>>2877330>LibreOffice has existed for decades yet all big businesses still subscribe to Microsoftit's not as clear cut because msft 365 also throws in teams, sharepoint, one drive and active directory. which is effectively replacing dropbox, slack, and an IT management service/LDAP setup. also libreoffice is kind of shit since everyone is using their browser and expect live document editing, the closest libre alternative is nextcloud or opencloud or whatever, which is self-managed AKA you need to hire someone to support it in-premises. msft 365 is a no-brainer really.
>So the idea that a Chinese model will "breach containment"The "idea" is that AI startups that are downstream from ChatGPT or Anthropic suddenly have a future beyond waiting until either jacks up the price, since they can just easily rebrand a chinese model and offer it as some sort of western alternative, as i explaned below, which is why seed investors are being pitted against SPVs heavily invested into frontier labs. Simply put you cant have open weight models and frontier labs co-exist in the same space for long, it's either one or the other.
>>2877371>The "idea" is that AI startups that are downstream from ChatGPT or Anthropic suddenly have a future beyond waiting until either jacks up the price, since they can just easily rebrand a chinese model and offer it as some sort of western alternative, as i explaned below, which is why seed investors are being pitted against SPVs heavily invested into frontier labs. Simply put you cant have open weight models and frontier labs co-exist in the same space for long, it's either one or the other.not the anon. please translate this into retard.
>>2877377companies like cursor and lovable just wrap access to chatgpt or claude in a pretty interface, so the minute chatgpt or anthropic start raising prices, which is what they intend to do to make their investors whole, they either absorb the rising costs and go bust, or pass the cost to consumers and go bust. meaning that their existence literally depends on what AI providers decide their price should be. with open weight models, they can just grab kimi, retrain it a little bit, and pass it as their own model, and not be tied to the whims of either chatgpt or anthropic. which is what cursor did with composer 2.
>>2877379 (me)
open weight models also mean that openai and anthropic effectively lose the defacto monopoly over the AI space, which means that they can no longer raise prices as a cartel to recoup investor money because consumers will just go to another provider. so if they have no avenue for profit through cartelized pricing schemes, then they can just go bust. so you see, either open weight models are allowed in the US, and companies like cursor live another day while openai dies, or open weight models are banned, cursor dies, and openai and anthropic get to dominate through a government-mandated oligopoly.
>>2872811who is this moron anyway?
>a Visiting Fellow at Heritage Foundationah yes…
>served as Senior Policy Advisor for Artificial Intelligence and Emerging Technology at the White House Office of Science and Technology Policy>was the primary staff drafter of America’s AI Action Planokay
>he also serves as a Visiting Lecturer at Yale Law School>where he co-teaches the Colloquium on Frontier AI Governancewhatever that means
<a lot more yada yada about AI and whatever policy>graduated magna cum laude from Hamilton College in 2014 with a B.A. in HistoryLMAO. why is it the most unqualified people lobby for and make policy for shit they have zero qualification to talk about in any way?
>>2877342>At the end of the day capital will embrace whatever reduces socially necessary labor time to produce commodities, which means embracing whatever is most efficient.But I thought we lived in the age of imperialism and monopoly capitalism. Only losers have to compete.
Unique IPs: 16