With all the people saying that you're going to have problems because the LLM is not good at refactoring or large code bases or OOP, etc. the point may be that if you're working to develop skills, an LLM herder might be a good one. Even if the models are almost, but not quite, good enough yet - they will be.
When looking for a career move you might want to focus on the trajectory more than the current state.
this is the pitch - it's open source, run it yourself. But >99% of people will not have the hardware needed to run these models at a high enough quality to be close to SOTA. So they will run the open-source models on CCP systems for a good price.
There is not a single LLM provider I trust enough to send secrets to. If you firewall accordingly the provider (or local) can be interchangable, barring capability differences of course.
I also struggle to find a provider that can credibly convince me I wouldn't be a product for when using. Have you found one?
Unless you're very careful, it's trivial to have my secrets to be sent to the LLM. If it reads your .env just to see the variable names, the secrets have been sent to the servers. Now - they probably don't care about you and your secrets - but it makes me more uncomfortable that they have them.
This is true of anthropic or openai - but for some reason I think the us govt or anyone else will have a harder time getting to my data from them than the CCP will any chinese company.
> but for some reason I think the us govt or anyone else will have a harder time getting to my data from them than the CCP will any chinese company.
US tech companies voluntarily give their data to the US government. Don't you remember PRISM? You think they stopped doing that?
> Internal NSA presentation slides included in the various media disclosures show that the NSA could unilaterally access data and perform "extensive, in-depth surveillance on live communications and stored information" with examples including email, video and voice chat, videos, photos, voice-over-IP chats (such as Skype), file transfers, and social networking details.[2] Snowden summarized that "in general, the reality is this: if an NSA, FBI, CIA, DIA, etc. analyst has access to query raw SIGINT [signals intelligence] databases, they can enter and get results for anything they want."[13]
> for some reason I think the us govt or anyone else will have a harder time getting to my data
US companies are required by law to hand over your data if given a warrant by USG. They don't need a warrant if they have a subpoena for less invasive data, or a FISA request. They can also ask without any justification, and see if the company will cough it up anyway (they often do). Any AI company with government contracts will want to give up data quicker so as not to threaten deals worth hundreds of millions.
> but for some reason I think the us govt or anyone else will have a harder time getting to my data from them than the CCP will any chinese company
With what we know about the US government's mass surveillance in cooperation with tech giants, I would highly doubt that either country is "better" in this regard
>I think the us govt or anyone else will have a harder time getting to my data than the CCP will any chinese company.
Why? You dont think that 5 eyes cyber peeps use every advantage they can get? And on the way out leave a dusting of evidence pointing at the russkies or chinese?
Why would two companies burning 100s of billions of dollars and are not profitable be safe keepers of your data when there is a huge market for all of that in the us and the us has really weak protections for those things so the companies can sell it to defense agencies?
Thing is, either way your data is getting hoovered up. If not today then eventually. It's just a matter of where. If you work in an industry where nation states might want to do you irreparable harm then yea don't let your data leave the country.
> If it reads your .env just to see the variable names, the secrets have been sent to the servers.
But these are development secrets. They're of no use to anyone.
Production secrets probably live in another repo, or at least are stored encrypted. In my case both. And I'm managing the least sensitive stuff you can imagine: a pre-school website, a non-profit event website, personal blog...
It's unlikely that you're special enough that someone will genuinely look through the massive amount of data produced by this system in order to target You Specifically. If you are that special you can just use another provider.
From this line of reasoning, my guess is that the huge discount is not so much intended to sell the data collection system as much as it is intended to sell the model. If you had to wring a geopolitical consequence from this, it would be that the US labs producing models would be impacted by a vastly less expensive competitor.
Are you actually planning on travelling out of the country right now? It's probably not a good idea even if you don't use Chinese products, which by the way you definitely do.
The people that travel out of the country are typically not the same ones aligned with the current administration. The vast majority of the MAGA base are more likely to not have a passport, while a large portion have probably never left their state.
I don't think anyone will ever be confused by the "only country to use the bomb" in this context. Your pedantry is not something to not feel bad about as it does nothing constructive to the conversation
only different if you don't consider "the bomb" the way it is colloquially used. until today, i would have never considered saying "the bomb" would not have been understood as a nuclear weapon whether that was an a-bomb or an h-bomb.
I mean I can't believe I have to say this explicitly but it should be assumed that any data you send to China can and will be used against our interest by the CCP...
Kicking myself when my little vibe coded widget to notify me when socks go on sale that does not and never has functioned properly is wielded as a mighty scepter to topple western hegemony
reply