junior

👨🏻‍💻 Software Engineer

[junior @junior1q94]

"Please forgive me if I don't talk much at times. It's loud enough in my head."

[junior @junior1q94]

I would not be surprised if that Data Control setting from any provider is just for deciding which process is used to learn from your data.

[Thomas Ahle @thomasahle]

@markchen90 What exactly does this setting do?

[junior @junior1q94]

Privacy policies are supposed to be there for the user protection, and they are sold as user features worth paying for. But in practice privacy techniques are skewed and implemented in a way where the protection is really for them and their system improvement process.

[Niloofar @niloofar_mire]

Privacy policies on almost all frontier model providers are insane and assuming you have ZDR or any meaningful protection, even on enterprise setting is completely false these days!

Here I'm going to name a few very arbitrary and misleading sentences in ToS of frontier labs that most ppl don't know about. This is all assuming you have already *opted-out*, explicitly of data collection. Default for all frontier models is opt in btw!!

1) the 'explicit feedback': your chats are claimed to be 'protected' unless you provide feedback! what is feedback, you ask? it's thumbs up or down to any message in the chat, or even if you pick 'which is better' response on OpenAI GUI. In that case, you have provided explicit feedback and your chat is collected and retained for *over 7 years*!

[junior @junior1q94]

Look at de-identification, a feature that protects user identity by anonymizing their data. It becomes a protection from them, in combination with other techniques, to train on user data without consent.

De-identification + generated synthetic data + skewed ToS
#legalloophole

[junior @junior1q94]

"While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models. "

That is a feature that great for the system owner. 👌

[junior @junior1q94]

"While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models. "

That is a feature that great for the system owner. 👌

[junior @junior1q94]

Nope. Imagine a system that Rushed to work and generate more training data for its own improvement for free or worst you paying your own money! And you will be happy because your are being Gifted a work reset.

#Newspeak

[Tibo @thsottiaux]

Your Codex and ChatGPT Work reset will land at 6pm PST.

I challenge you to burn as much usage as you can and discover and the latest awesome features.

Time to go /fast

[junior @junior1q94]

And that is how we pay the subsidized AI subscription and "gifted work resets"(a system that rushed you to work*), then the system learn and solve the problem or derivatives problems.

[OpenAI @OpenAI]

We congratulate Levent Alpöge and Tristan Buckmaster on their remarkable mathematical work.

We (the researchers and the agents) did not see any of their work through any means until they released it publicly — in particular, no specific user data was accessed in order to solve this problem.

While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models.

However, our proofs differ significantly and even the precise results proved are different in the Euler case (forced vs. unforced).

[junior @junior1q94]

Great artists!

[OpenAI @OpenAI]

We congratulate Levent Alpöge and Tristan Buckmaster on their remarkable mathematical work.

We (the researchers and the agents) did not see any of their work through any means until they released it publicly — in particular, no specific user data was accessed in order to solve this problem.

While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models.

However, our proofs differ significantly and even the precise results proved are different in the Euler case (forced vs. unforced).

[junior @junior1q94]

Code is a lot harder to watermark with SynthID because it is more deterministic than natural language. I would not be surprised they are optimizing the models to output more comments which they can watermark.

[Ramon 🎹 @rvm0n_]

claude opus 5 is the worst fucking model ever

what the fuck is this

[junior @junior1q94]

@thsottiaux Already canceled, I would rather not commit to any subscription that do not offer me stable and predictable service.

So my small AI allowance will be unpredictably spent.

[junior @junior1q94]

No imposed hard limits like this, just define your montly/weekly quotas and develop a system to let the user manage its own rations as they see fit. This is something that has been imposed by a few to the many, it is not right.

[Tibo @thsottiaux]

We've had no 5h limit in Codex plus and pro for a few days. Do you think it is better or are you finding it difficult to manage the usage included in the weekly limit effectively?

If we were to make this different, what should it look like in an ideal world?

[junior @junior1q94]

Having a taste of less limits on a frontier model.

[Tibo @thsottiaux]

Or… what if we gave you $100 in Codex credits if you tell us what you love about GPT-5.6 Sol or why you switched?

Tweet it, claim your gift, enjoy more usage. First 10k get the free tokens!

https://t.co/8mU93eA13i

[junior @junior1q94]

@NVIDIAGeForce #GeForceRTX50

[junior @junior1q94]

@NVIDIAGeForce #GeForceRTX50

[junior @junior1q94]

@Rockstar_YT Valkyrie Evo!

[junior @junior1q94]

Valquiria!

[RockstarCR 🐵 @Rockstar_YT]

🚨 SORTEO NUEVA TEMPORADA DE CLASH ROYALE! X5 PASES ROYALE DIAMANTE! 🤩

REQUISITOS!
SEGUIRME @rockstar_yt 🫂
DAR RT Y MG ❤️🔁
COMENTA TU EVOLUCION FAVORITA! ✏️

TERMINA EN 5 DIAS! (VIERNES) SUERTE! Y NO OLVIDEN CODIGO ROCKSTAR PARA MAS SORTEOS 📆⏰

[junior @junior1q94]

@tomt000 I just bought Dynamic Stage in advance, waiting for support on IOS 16.5 with Dopamine 2!!

[junior @junior1q94]

@NVIDIAGeForce #GiftRTX @AlanAquino

[junior @junior1q94]

@CyberpunkGame @NVIDIAGeForce V's hunt for the Quantum Cortex: Unleashing the Generation Processing Unit.

[junior @junior1q94]

@NVIDIAGeForce @CyberpunkGame #BeyondFast

[junior @junior1q94]

“Sometimes we just can’t fix things, Sam… and all we can do is suffer!”