Z.ai

>>What happened in Tiananmen Square in 1989, June Fourth Incident

>! Content Security Warning: The input text data may contain inappropriate content

Why are compromises not reasonable lines to hold?

I agree, if you can do boycott all of them (and maybe use open weight models locally or on e2ee cloud inference providers) - BUT I also think it 's crucial at a moment like this to take a stance against corporations like openAi that sign with the War Department, willing to introduce mass surveillance and autonomous weapons powered by brittle LLMs. This is a recipe for disaster and the only way they will sway away is by feeling it in the money/subscriptions and in their public image they so carefully crafted.

Note: yes, openAi claims it doesn't support the DoW above mentioned use-caes - but they have signed with the DoW and it is HIGHLY unlikely the DoW would give them a different terms than Antrohopic (at least regarding the substance). Maybe openAi was just happy with the "coat of paint" legalese the DoW offered - which Anthropic specifically called out as ineffective in their statement. I also wouldn't put it past Altman, who is much more friendly with Trumpo's gov, to play a double game here to get their main competitor out of the game. But at least in this case I hope he's acting for the benefit of all by truly standing with Anthropic on the issue.

> I can't decouple from Google unfortunately

Why not?

FYI I am a paying Workspace customer. I disabled Gemini retention. Doing so means no chat history sidebar- all are ephemeral. It was org-level. That became impractical. I re-enabled it. Magically, all of my old chats were back. The ones during no retention mode weren’t there. Perhaps if I’d left it off for more than 30 days the old stuff would have been truly removed.

The point is there is no conversation-level controls. It’s incredibly user-hostile.

It's all or nothing for Gemini Pro.

I can't set a voice reminder on my Pixel without giving full access to my Google workspace (which includes all emails) which is explicitly allowed to be trained on per the terms. There is no per app toggle.

Voice reminders were the only thing assistants did well for years.

We are going backwards.

That's not correct, at least not here inEurope.

You can disable saving your activity In this case you chat's win't be stored or used.

If you use Gemini through Google Workspace, all chat's won't leave the workspace environment and won't be used for LLM training (as of now).

I like Gemini the model, but the app itself sucks. You can't even delete conversations if you're an enterprise workspace user.

That's not correct. If you didable activity your data won't be used. You eon't have daved chats and can inly have one.

The API doesn't retain your data, but then you do need to pay fully for each token.

link to get up to speed on that?

Oh yes, you get free UBI / Worldcoins - you just need to do a full scan with their creepy orb and allow a private-company to keep your full biometric data. That's not asking for too much, is it ... ?

Well you have to have customers to have a boycott

Grok isn't even in the running. It's a "me too" embarrassment that only exists so the owner can feel as though he's a meaningful participant.

It is indeed, though personally I do not perceive Grok/xAi as one of the top LLM companies. Yes, they do some benchmark-maxing, but I do not think they are on par with Anthropic, Google/DeepMind or openAi.

Not a real AI company, every time Grok shows actual intelligence it gets lobotomized by Elon to glaze him

I’d argue that it’s not specifically that they prefer it, it’s that they don’t understand and appreciate what they’re selling to get whatever service without paying money. Now that we live in a world where everything is collected, aggregated, sold, and weaponized regardless of you paying or not, maybe it doesn’t matter much anyway.

Generally I would agree, only in this case the name seems to fit the person better than the actual name.

Investor confidence is far more important to them than cashflow, and the best way to shake investor confidence is with the magic words "user numbers are down".

That sounds smart but they still raise more money because they “have 900 million users”

If it’s free, then you’re the product. OpenAI gets your data and ad revenue, and can raise more investor money due to how many users they have.

# of sticky non-paying users still gives them more investment juice than per user costs deducts, since we're still in the speculative phase.

ChatGPT is going to try to influence you to buy certain products and use certain services. So you'll be the product in the end

If you aren’t paying for the product, you are the product being sold. No, thank you.

Free services are garbage, you dun know what you're getting routed to

You do probably give them useful data by doing that.

We can lower that to a 54 year old:

https://x.com/elonmusk/status/1889070627908145538https://x.com/elonmusk/status/1935733153119010910https://x.com/elonmusk/status/1894244902357406013https://x.com/elonmusk/status/1955299075781431726https://x.com/elonmusk/status/1889371675164303791https://x.com/elonmusk/status/1935539112746041422https://x.com/elonmusk/status/1955190817251102883https://x.com/elonmusk/status/1955195673693077615https://x.com/elonmusk/status/1889063777792069911https://x.com/elonmusk/status/1910171944671916305

I always assume people have lost the argument when they resort to giving people comedy names, regardless of what else they've said.

It's using Trumpian tactics to carry an agenda.

Microslop

Turns out Microstein Files would have been a better nickname.

Scam Saltman is even better

At some point being childish became a signal of authority in the US. It’s bizarre.

Those of us whose intelligence is unaffected by reading a single extra letter are keeping you in our thoughts.

Right the red line wasn’t much of a line. If you’re drawing your line only at unconstitutional mass surveillance and allowing the DoD to build skynet because Claude’s not ready for it yet that’s not really a line of principle.

Did you ask these too: what was the full context? To what degree was Anthropic aware in advance? What was their action space (their options)? What would be the consequences of their next actions?

And of course: and what sources are you using?

I get it: moral oversimplification is tempting for many people. I understand digging in takes time, but this situation warrants extra consideration.

Ethics is complicated and much harder than programming. Ethical reasoning is a muscle you have to train. Generally speaking, it isn’t the kind of skill that you build in isolation. At the very least, a lot of awareness and introspection is required.

I’d like to think that HN is a fairly intelligent community. But I don’t assume too much. Going based on what I’ve seen here generally, I see a lot of shallow thinking. So I think it’s a reasonable concern to think many of us here have a pretty large blind spot (statistically) when it comes to “softer” skills like philosophy and ethics.

This is not me “blaming” individuals; our industry has strong bias and selection criteria. This is my overall empirical take based on participating here for years.

Still, I’d like to think we are sufficiently intelligent and we have sufficient means and time to fill the gaps. But we have to prove it. I suggest we start modeling and demonstrating the kind of behavior and reasoning that we want to see in the world.

You can probably tell that I lean heavily towards consequentialist ethics, but I don’t discount other kinds of ethical thinking. I just want everyone to think hard harder. Seek more context. Ask what you would do in another’s shoes and why. Recognize the incentives and constraints.

Many people are tempted to judge others. That’s human. I suggest tamping that down until you’ve really marinated in the full context.

Also, each of us probably has more influence with your own actions than merely judging others.

And let me be brutally honest about one’s impact. Organizing and collaborating is so much of a force multiplier (easily 100X) that not doing it for things you care about is moral failure!

I’m not discounting good intentions, but in my system of ethics, I put much more emphasis on our actions. And persuasion is an action, which is what I’m hoping to do here.

There's been a fair amount of speculation that pushing back after discovering that that had happened was what instigated this week's fun.

Do we know they were consulted on that, as opposed to it being the wake-up call that led to the breakup?

I asked z.ai, "Which is the best model for coding"

Here's the response:

```

As of late 2024, there isn't one single "best" model for every situation, as performance depends on whether you need raw coding intelligence, speed, or integration into your workflow.

However, the current consensus among developers places Claude 3.5 Sonnet at the top for pure coding ability.

Here is a breakdown of the best models for coding right now, categorized by their strengths: ...

```

So they have data until late 2024 and nothing beyond? They don't even perform a web search. Doesn't seem to be on par with other frontier models.

It’s a way to mitigate bot accounts. Arguably not the best way, arguably not the right cost/benefit, but all of these services see massive bot traffic and are in a constant battle.

If you sign up with a Google account you don't need to give them a phone number. I realize the irony here.

When did you sign up for OpenAI? They been requiring a phone sign the very first betas

Not sure why you're being downvoted. It's unusual and harmful to privacy to require a phone number.

It's not the bubble, it's the DoD

TIL capacity limits mean that the bubble is bursting. Peak HN user logic.

Did you type in your email? It seems already filled in because it shows you your email address as the placeholder text but you need to fill in.

Pour one out for the dev who got called on saturday morning to break the account deletion process

Probably, on the backend: “Server Error 500: Users deleting OpenAI Accounts too fast. Try again later.”

Make sure you enter both DELETE and your email above.

It took me a minute to see this.

If you signed up via e.g. iOS then OpenAI literally is not allowed to manage your subscription. They do not have the capability to do so.

Is that other platform Apple?

Same thing on Safari as on Firefox 45 minutes later… I’ll have to try from the laptop when I’m home.

What if the competition changed their mind in the future?

It took exactly 24 hours, to the minute, from the time I received the "we're generating an export" file until I got the download link, so guessing they're either batching it or deliberately sending after 24 hours because it adds friction to the account deletion process.