AI: Privacy Guarantor asks DeepSeek for information
Possible risk to the data of millions of people in Italy
The Guarantor for the protection of personal data has sent a request for information to Hangzhou DeepSeek Artificial Intelligence and Beijing DeepSeek Artificial Intelligence, the companies that provide the DeepSeek chatbot service, both on the web platform and on the App.
The Authority, considering the potential high risk for the data of millions of people in Italy, has asked the two companies and their affiliates to confirm which personal data are collected, from which sources, for which purposes, what is the legal basis of the processing, and whether they are stored on servers located in China.
The Guarantor also asked companies what kind of information is used to train the artificial intelligence system and, in the event that personal data is collected through web scraping activities, to clarify how users registered and those not registered to the service have been or are informed about the processing of their data.
Within 20 days, companies must provide the Authority with the requested information.
The model works, sure, but that's not the point. OpenAI and others can replicate any technical advances for future models moving forward. DeepSeek R1 will be "outdated" in 6-12 months and be replaced by other, more powerful models for sure. If DeepSeek's success is based on stolen data, they won't be able to replicate their success as for sure such security gaps will now get closed.
There have been talks about a laissez-faire attitude regarding cybersecurity at OpenAI for a long time.... but this is surely coming to an end now. Same at Google.
I want to believe DeepSeek R1 is legit… but the more details emerge, the more it feels like something isn’t right.
The claim that R1 was trained for under $6M on 2,048 H800 GPUs always seemed suspicious. Efficient training techniques can cut costs, sure—but when OpenAI, Google, and Meta are all burning hundreds of millions to reach similar benchmarks, it’s hard to accept that DeepSeek did it for pennies on the dollar. Then Alexandr Wang casually drops that they actually have 50,000 H100 GPUs… what happened to that “low-cost” narrative? If this is true, it's not efficiency—it’s just access to massive hidden compute.
The stolen OpenAI data theory is another red flag. OpenAI researchers have been hit by multiple security breaches in the last few years, and now we have a former OpenAI engineer found dead under very weird circumstances. Coincidence? Maybe. But corporate espionage in AI isn’t some sci-fi plot—it’s very real, and China has been caught running large-scale operations before (Google exfiltration cases, the ASML trade secret theft, etc.).
And then there’s the CCP-backed propaganda angle. This part is almost too predictable—China hypes up a “homegrown” breakthrough, gets state media to push it as “proof” they’ve surpassed the West, then quietly blocks foreign scrutiny. Lei pointed out that DeepSeek won’t even let U.S. phone numbers register. Why? If R1 is truly open-source and transparent, why limit access? We’ve seen this before with ByteDance, Alibaba, etc.—government-approved success stories that follow a controlled narrative.
But despite all that skepticism… R1 is real, and the performance numbers do exist. Whether they’re running stolen training data or smuggled GPUs, they’ve built something that competes with OpenAI’s o1. That’s still impressive. The question is how much of this is a real technological leap vs. how much is state-backed positioning and/or cutting corners.
So what happens next?
If DeepSeek is serious, they need outside audits—actual transparency, full datasets, external verification. Not just another “trust us” moment.
The U.S. needs better export control enforcement… we’re seeing massive loopholes if China can stockpile 50K H100s despite all the restrictions.
AI labs (OpenAI, Anthropic, etc.) need better security. If OpenAI’s data really did leak, this won’t be the last time.
I don’t think R1 itself is a scam, but the surrounding story feels curated, opaque, and suspiciously convenient. Maybe DeepSeek has built something remarkable, but until they open the books, I can’t take their claims at face value.
GitHub Copilot Edits (Preview) in Visual Studio 2022 combines the conversational flow of chat and an inline review experience to help you iterate across your codebase with more control and efficiency.
The OpenAI vs. DeepSeek debate is fascinating... but I think people are oversimplifying both the challenges and the opportunities here.
First, OpenAI’s valuation is a bit wild—$157B on 13.5x forward revenue? That’s Meta/Facebook-level multiples at IPO, and OpenAI’s economics don’t scale the same way. Generative AI costs grow with usage, and compute isn’t getting cheaper fast enough to balance that out. Throw in the $6B+ infrastructure spend for 2025, and yeah, there’s a lot of financial risk. But that said... their growth is still insane. $300M monthly revenue by late 2023? That’s the kind of user adoption that others dream about, even if the profits aren’t there yet.
Now, the “no moat” argument... sure, DeepSeek showed us what’s possible on a budget, but let’s not pretend OpenAI is standing still. These open-source innovations (DeepSeek included) still build on years of foundational work by OpenAI, Google, and Meta. And while open models are narrowing the gap, it’s the ecosystem that wins long-term. Think Linux vs. proprietary Unix. OpenAI is like Microsoft here—if they play it right, they don’t need to have the best models; they need to be the default toolset for businesses and developers. (Also, let’s not forget how hard it is to maintain consistency and reliability at OpenAI’s scale—DeepSeek isn’t running 10M paying users yet.)
That said... I get the doubts. If your competitors can offer “good enough” models for free or dirt cheap, how do you justify charging $44/month (or whatever)? The killer app for AI might not even look like ChatGPT—Cursor, for example, has been far more useful for me at work. OpenAI needs to think beyond just being a platform or consumer product and figure out how to integrate AI into industry workflows in a way that really adds value. Otherwise, someone else will take that pie.
One thing OpenAI could do better? Focus on edge AI or lightweight models. DeepSeek already showed us that efficient, local models can challenge the hyperscaler approach. Why not explore something like “ChatGPT Lite” for mobile devices or edge environments? This could open new markets, especially in areas where high latency or data privacy is a concern.
Finally... the open-source thing. OpenAI’s “open” branding feels increasingly ironic, and it’s creating a trust gap. What if they flipped the script and started contributing more to the open-source ecosystem? It might look counterintuitive, but being seen as a collaborator could soften some of the backlash and even boost adoption indirectly.
OpenAI is still the frontrunner, but the path ahead isn’t clear-cut. They need to address their cost structure, competition from open models, and what comes after ChatGPT. If they don’t adapt quickly, they risk becoming Yahoo in a Google world. But if they pivot smartly—edge AI, better B2B integrations, maybe even some open-source goodwill—they still have the potential to lead this space.
An 80% size reduction is no joke, and the fact that the 1.58-bit version runs on dual H100s at 140 tokens/s is kind of mind-blowing. That said, I’m still skeptical about how practical this really is for most people. Like, yeah, you can run it on 24GB VRAM or even with just 20GB RAM, but "slow" is an understatement—those speeds would make even the most patient person throw their hands up.
And then there’s the whole repetition issue. Infinite loops with "Pygame’s Pygame’s Pygame’s" kind of defeats the point of quantization if you ask me. Sure, the authors have fixes like adjusting the KV cache or using min_p, but doesn’t that just patch a symptom rather than solve the actual problem? A fried model is still fried, even if it stops repeating itself.
On the flip side, I love that they’re making this accessible on Hugging Face... and the dynamic quantization approach is pretty brilliant. Using 1.58-bit for MoEs and leaving sensitive layers like down_proj at higher precision—super clever. Feels like they’re squeezing every last drop of juice out of the architecture, which is awesome for smaller teams who can’t afford OpenAI-scale hardware.
"accessible" still comes with an asterisk. Like, I get that shared memory architectures like a 192GB Mac Ultra are a big deal, but who’s dropping $6,000+ on that setup? For that price, I’d rather build a rig with used 3090s and get way more bang for my buck (though, yeah, it’d be a power hog). Cool tech—no doubt—but the practicality is still up for debate. Guess we'll see if the next-gen models can address some of these trade-offs.
Thanks.. what I am interested in is the the type of censorship that has been applied and what effects prompt-jailbreaks have. Personally I have limited experience with that
DeepSeek is a censored product and by that of limited use for anything that might require prompts that are about anything that is somehow "controversial" in the eyes of the CCP. However, the censorship seems to be applied for certain prompts and doesn't seem to be integrated into the model itself as the answers given to such prompts are very similar and generic.
Has anybody already been able to successfully use prompt jailbreaking or other tricks to overcome this? It would be interesting to see what DeepSeek actually knows instead of what it is responding.
Censoring a model via selective training data or post-training is much more difficult.
The possible "solutions" applied to this "problem" (in the eyes of the censors) will be of high importance moving forward.
Other gov. actors also have an interest in altering models, let's not forget.
I get the point about conversations naturally splitting when there are 5+ people... definitely happens to me all the time. But I wonder if it’s really just about mental limits, or if there’s something else going on? Like, isn’t it also about the vibe of the group or the type of people involved? Some groups are just better at keeping everyone engaged no matter the size- like certain friend groups or teams that have great dynamics.
Also, not sure if the "four is magic" thing holds up everywhere... In my experience, some of the best conversations happen with just two people. Like really deep, meaningful stuff you can’t get with more people. And for bigger groups, there’s often this chaos energy that can be fun in its own way. Yeah, it’s not the same as an intimate chat, but it’s not worse, just different.
That said, I do like the idea that our brains are wired for certain sizes - makes sense when you think about the mental juggling it takes to track other people’s thoughts and reactions. And I love how Dunbar tied this to Shakespeare—kinda cool that he instinctively kept scenes small to avoid "cognitive overload." Makes me wonder if modern writers and creators even think about this stuff or just stumble into it.
So yeah, the mental limits idea is interesting, but I feel like the type of people, the setting, and even the purpose of the group matter a lot too. Sometimes it’s not the number of people, but how good they are at making everyone feel included... which is maybe a rarer skill than we think.
uv is cool, no doubt—super fast, love the idea of managing Python versions + deps all in one place. But VC funding... Rust... unofficial Python builds... lots of red flags here if you think about the long-term impact on the ecosystem.
VC-backed tools always make me a little nervous. Sure, Astral says the tools will stay free, and I get that they're targeting enterprise for revenue (private package registries, etc.). But how many times have we heard this before? “Don’t be evil,” right? We’ve seen companies pivot to paywalls or watered-down open-source tools after funding dries up. What guarantees do we have here that uv won't eventually fall into that same trap?
Yeah, Rust is amazing (ruff is a beast), but it’s still a small subset of devs compared to Python. If Astral folds or just loses interest, how easy is this really going to be for the Python community to maintain? Forkable? Sure. But forkable ≠ maintainable when the dev pool is tiny.
Also, what's with using unofficial Python builds by default? Even on platforms where official builds exist (macOS/Windows)? I get that these standalone builds solve certain problems (like bootstrapping), but it feels like a risky shortcut. If those unofficial builds go unsupported or change directions, where does that leave "uv"? Why not at least give users the option to rely on official binaries?
And fragmentation... Python tooling is already a mess. Pip, poetry, conda, pyenv, rye, flit, etc.—do we really need another tool in the mix? Feels like every new tool just promises to "fix packaging forever" but ends up adding another layer of complexity. Why not contribute these improvements to existing tools like pip? Sure, innovation is great, but at what point does it become too much choice and not enough cohesion?
uv looks great, and I love the speed + features. But the ecosystem-level risks here... hard to ignore. Would love to see some stronger guarantees around open-source sustainability, community governance, and alignment with Python standards before jumping in fully. Otherwise, it’s just one more shiny tool that could end up abandoned or locked behind a paywall in 5 years.
COVID hit hard, but these gaps were growing before the pandemic... especially for low-income and struggling students. Remote learning just made it worse. Socioeconomic status still drives most of this—wealthy kids have resources, while poorer kids are left behind. Add in demographic shifts—immigrant kids doing better here than at home but still starting from different baselines—and the system isn’t adapting fast enough.
Charter schools? They cherry-pick students and leave public schools with fewer resources and more challenges. Standardized tests like TIMSS highlight the problem but don’t explain why. Feels like a perfect storm: inequality, bad policies, underfunding, and no clear plan to fix any of it.
While these advancements in quantum error correction are undeniably impressive, I can't help but feel a sense of déjà vu here.... reminds me of the decades of "breakthroughs" in fusion energy—always promising to change the world, but perpetually just out of reach.
The Google benchmark with random circuit sampling is fascinating from a theoretical perspective, but it’s hard to see how this translates into solving problems that matter outside the quantum research community.
The lack of practical applications is a glaring issue. Sure, it's cool that Willow can outperform Frontier on this obscure task, but where’s the real-world impact?
Cryptography, optimization, drug discovery—these are the kinds of problems quantum computing needs to beat if it’s going to justify the investment. Until that happens, it feels like we’re stuck in a cycle of overpromising and underdelivering, with flashy press releases but no tangible results.
And let’s talk about scalability. Even if Willow hits the error-correction frontier, the number of physical qubits needed to build a truly practical quantum computer seems astronomical. Millions of qubits just to factor a number?
It’s hard to see how this scales in a way that makes economic or scientific sense. Right now, it feels like quantum computing is a field for researchers who are okay with not seeing practical outcomes in their lifetimes.
Maybe I’m too cynical, but this smells like another example of tech marketing getting ahead of the science. Maybe we can admit that we’re still decades away from quantum computing having any real-world relevance?