I spent a couple years working at the platform (AndroidTV + Google Assistant) level on smart TVs.

> any that want to do Alexa style voice prompts need it

Not technically true. More modern smart TVs (~post-2021) put speech recognition models for the hotword only onto the DSP, and then have the DSP wake up the CPU for further auditory processing when it detects a hotword or remote button press. There are major power-efficiency benefits to this, because the DSP uses a fraction of a watt and can easily keep running while the TV is sleeping, while the CPU, Internet, and rest of the circuitry is a major power hog. Your smart TV or streaming dongle would basically never sleep if it were forever running the audio capture.

For that matter, it's pretty rare for the device to actually run the speech recognition models on-device. Smart TV manufacturers are very sensitive to BOM costs, which means that their hardware is only powerful enough to run machine-learning models at the extreme high end of their product range. Most of the time the full audio gets streamed back to the Internet and the speech recognition & voice assistant gets run in the cloud, with commands streamed back to the TV.

I won't say that it never has always-on audio listening; a significant number of devices I tested did. But I'll say this comes because they're poorly implemented. The DSP hotwording approach requires coordination between hardware and software developers, oftentimes across companies, and it's more complex to implement and a lot of executives don't care about the marginal consumer benefit it gives. Many smart TVs just guzzle power when they're off instead.

Except it's not covered in their TOS

Why and what reference do you rec instead? Even this site gets fake news that stays up for hours.

Sure, ‘knee-jerk’ reddit is unreliable, but many upvotes over time with supporting comments can be a strong signal of truth

Why? Their system is literally mined for its Q&A merit by search engines. If it was all very bad information, NYSE: RDDT would not exist.

They still can’t technically obtain it without a warrant. They do it anyway.

Who bears responsibility when TV makers and data brokers undoubtedly had clear and verifiable evidence passing through their systems?

Follow the money

I think it's just a logical conclusion if everyone keeps feeding OpenAI and Anthropic data. I don't think it's imminent, but if AI keeps improving and regulation is very lax I could see this world developing in 5-10 years.

Say OpenAI builds GPT-11 and they realize it can do 95% of their clients work. Why not hire away key people from the client or the client's competitors and start eating into their market share? OpenAI would realize if GPT-11 has this ability then Anthropic is probably able to do the same thing. The company that has the most money and credit available is going to be able to buy more compute. OpenAI and Anthropic cannot slow down unless mandated to by the government.

Announcement

Another anecdotal data from an arxiv preprint [0]: "After the paper was written, the author learned that the same result was also proved independently and by a different method by Ningyi Li [7]. The result is also contained in a paper [8] released by OpenAI on October 6, 2026, in which the proof strategy and specific choices of notation are identical to a preliminary version of the present paper that was uploaded to ChatGPT on September 8, 2026."

So not only did OpenAI publish a proof similar to one discussed on ChatGPT, but also there were other possible strategies to prove the result, that OpenAI didn't follow.

Also, up to my knowledge, on the free plan, OpenAI doesn't claim that they won't use the chat data for training. So this seems plausible.

[0]: https://arxiv.org/abs/2610.10072

The "coincidence" defense never works for random individuals so I won't buy it when it comes companies worth billions which will lie, cheat, and deceive to make even more billions. They've been caught lying and stealing again and again. After so many years of this I think any trust put in them is either blissful ignorance or willful misdirection.

Claude is so absurdly bad about false triggers for security. It kept triggering when I asked it questions about how to plant some tree saplings I got because somehow that was a security risk.

Anthropic is a joke of a company.

I'm in the new CVP, and even Mythos gave me a cyber stop just for creating an MCP server that can remotely connect to my Steam Frame to run commands. It must've thought I was trying to create a centralized malware or something.

Not even sure why Mythos even still exists when it's identical to Fable in every way, including refusing to do stuff that looks like it could ever be repurposed into unapproved use cases.

We're all naked under our clothes, I don't see the problem. One of the funniest shock-bits I've ever seen was when Jon Stewart's book on American government and history included full-frontal (Photoshopped) nudes of SCOTUS, but I guess he couldn't get away with that during the current era of Puritanical panic.

It's just an adult human body, Michael, what could it cost? One('s) dignity?

Tastelessness aside, this seems perfectly rational to me. Unlike reverse engineering for the purpose of developing security exploits, "nudification" doesn't meaningfully expand the capabilities of malicious actors with access to free online porn and Photoshop.

The headline is misleading, but the "article" is only five screenshots of Twitter posts. Someone used AI to decrypt the Samsung TV's firmware and found always on listening, classification, lots of privacy nightmares. Then their (the hacker's, not Samsung's) OpenAI got suspended for "Cyber Abuse".

I agree with you, but if OpenAI wasn't suspending users for exposing cyber-spying that these consumer companies like Samsung are doing, more and more people would expose such issues.

Did you mean Mimo v2.6 Pro? Because Kimi K3.1 Pro is proprietary, and not open.

Also leading the frontier is much much harder than catching up to frontier.

Iirc They specifically trained them to not be useful in cyber stuff. But the point still holds

… IF chinese models stay open.

In fairness, they have not earned the benefit of the doubt based on past behavior.

Until Kevin quits and you need to modify his work

It is great if you have enough RAM. I have enough to run it, but not enough for prefix cache. 27b is so much faster with long term tasks.

The Qwen3.8-Flash-Next model is ok for reverse engineering. It can definitely solve simple tasks, but it gets easily confused, and burns tokens like crazy.

I've been sick with stomach flu the past few days, so I put together a small benchmark for all the top AI models where I ask them to reverse engineer a very simple small crackme (https://www.crackmes.one/crackme/6ab4ce5195b976f8f1300b54) using Ghidrasql (https://github.com/0xeb/ghidrasql)

(As an aside: I think representing tools for AI as virtual SQL tables is much better than traditional MCP servers, and that this approach deserves a lot more attention. Props to Elias Bachaalany for thinking out of the box and putting it together).

The results of the benchmark are below. It's sorted by cost, but I added notes about each AI model for myself. The notes try to talk about intelligence, speed, and cost. Intelligence is a subjective measure which I assign to each model by reading its thinking tokens and watching its tool calls. You can definitely tell which models are more intelligent, even on this small benchmarks, by their tool use, false starts, and how easily they fall for the simple misdirection in the crackme.

(gist is more readable on mobile) https://gist.github.com/markoglasgow/8812c19b6fbffed30b2661b...

```

  +----------------------------------+--------------------------------------------------------------------------------+-----------------+-----------------+-----------------+
  | Model / route                    | Experience / outcome                                                           |    Total tokens |   Output tokens | Cost            |
  +----------------------------------+--------------------------------------------------------------------------------+-----------------+-----------------+-----------------+
  | Xiaomi MiMo v2.6 Flash           | Incredibly fast and intelligent; solved.                                       |         225,291 |           3,281 | $0.005 (0.5c)   |
  | GLM 5.3 Flash / OpenRouter       | Least intelligent group but cheapest; solved.                                  |         340,006 |          11,205 | $0.01           |
  | ChatGPT Luna Latest              | Very fast, concise, token-efficient, intelligent, and cheap; solved.           |         401,837 |           5,119 | $0.01           |
  | Qwen 3.8 Flash                   | Initially confused and failed, then eventually solved.                         |        ~418,000 |          16,676 | $0.02           |
  | Qwen 3.8 27B                     | Fast; one wrong answer, then quickly self-corrected.                           |        ~774,784 |          12,658 | $0.02           |
  | DeepSeek Flash / direct API      | Fast, cheap, and accurate; solved.                                             |         369,334 |           9,089 | $0.03           |
  | DeepSeek Flash / OpenRouter      | Fast and intelligent; solved in 76 seconds.                                    |         371,689 |          13,888 | $0.03           |
  | Claude Haiku 5.5                 | Fast and surprisingly intelligent, but overly thorough and token-heavy.        |         662,346 |          26,244 | $0.03           |
  | Thinking Machines Inkling        | Low intelligence; good speed and relatively token-efficient; may loop.         |             N/A |             N/A | $0.03           |
  | Xiaomi MiMo v2.6 Pro             | Overthought the task and wandered into tangents.                               |        ~420,644 |          13,638 | $0.05           |
  | GLM 5.3 / OpenRouter             | Cheap and intelligent but behind DeepSeek; solved in about 2m 20s.             |       ~702,473* |          11,879 | $0.08           |
  | GPT Sol Latest (Sol 6.1)         | Slow, but highly intelligent and extremely token-efficient; solved in ~89s.    |         294,957 |           1,406 | $0.15           |
  | MiniMax M3                       | Least intelligent comparison; many wrong turns and repeated small mistakes.    |      4,546,552* |         116,804 | $0.20           |
  | GPT Terra Latest                 | Very fast, very smart, and token-efficient; solved.                            |         229,888 |           2,342 | $0.20           |
  | Mistral Large 4.0                | Medium speed; small mistakes but reached the solution.                         |         303,687 |           4,041 | $0.21           |
  | Kimi K2.7 Code                   | Less intelligent, slow, expensive; not worth using.                            |         100,740 |           8,938 | $0.22           |
  | Claude Opus 5.5                  | Highest intelligence and token efficiency; pretty fast; cheaper than Sonnet.   |         312,837 |           1,862 | $0.24           |
  | Qwen 3.8 Max 0902                | Fast and intelligent, better than Kimi K3, but token-hungry.                   |       1,305,241 |          26,390 | $0.25           |
  | Kimi K3 / OpenRouter             | Faster and cheaper than direct K3, but still very expensive.                   |         750,064 |           5,922 | ~$0.32          |
  | Meta Muse Spark Open             | Intelligent and medium speed, but among the more expensive runs.               |         872,850 |          10,284 | $0.35           |
  | Gemini 4 Argon                   | Extremely fast and intelligent; strong value for capability.                   |         384,046 |           2,870 | $0.35           |
  | Grok 4.7                         | Very fast and intelligent, but expensive.                                      |         655,648 |           4,217 | $0.46           |
  | Meta Muse Glimmer 30B            | Fast but low intelligence and token-inefficient; cancelled unsolved.           |             N/A |             N/A | ~$0.50 wasted   |
  | Claude Sonnet 5.5                | Smart and fast but very token-inefficient and expensive.                       |      1,243,285* |          20,381 | $0.57           |
  | Kimi K3 / direct API             | Intelligent but not notable; extremely slow and far too expensive.             |         205,710 |           3,215 | $0.72           |
  | ChatGPT Astra Latest             | Medium speed, token-efficient, highly intelligent; cost similar to Fable.      |         278,603 |           1,536 | $0.83           |
  | Claude Fable 5.1                 | OK speed, very expensive, medium token efficiency; solved in ~53s.             |        397,806* |           2,857 | $0.85           |
  | Gemini Flash                     | Refused the reverse-engineering task.                                          |             N/A |             N/A | Not reported    |
  | Thinking Machines Inkling Small  | Got stuck immediately and was terminated.                                      |             N/A |             N/A | Not reported    |
  +----------------------------------+--------------------------------------------------------------------------------+-----------------+-----------------+-----------------+

The latest Deepseek-Flash model was the most impressive of the open models. Intelligent, fast, cheap. What more do you want? I think the Xiaomi MiMo - Flash model did well, but it might be a fluke, given how poorly the Xiaomi MiMo Pro model performed (it wandered off on a million side quests).

What's also interesting is how competitive Luna is to the lower end Chinese models. It's quite a bit more intelligent, burns a lot less tokens, is plenty fast and very cheap. After doing this experiment, I don't think it's a foregone conclusion that the American AI labs will be having the bottom of the market eaten away by the Chinese models. It seems they (at least OpenAI) can compete on the bottom end as well.

Finally, the benchmark's results reinforce a few things people have been saying about Sonnet and Terra, namely, they're not competitive for a range of tasks. For easier tasks, Haiku/Luna are plenty capable. For harder tasks, Opus and Sol are so much more intelligent, that they can get them done quicker and cheaper. Maybe there is a place for them in the "Opus designs and Sonnet implements" workflows, but in that case, why not "Opus designs and Deepseek implements?"

PS Code harness used was Claude Code. $20 in openrouter credits was used to access the models. Kimi K3 and Deepseek were also accessed directly via the API and then compared to Openrouter. Kimi K3 specifically was much faster on Openrouter, and the difference in output tokens to get the same results for Deepseek/Kimi makes me think the Openrouter providers are hosting significantly quantized versions of the model. Goes to show that Moonshot.ai's acceptance testing of providers is a useful thing, otherwise your model might look artificially bad when it's hosted as heavily quantized by providers who dont care about quality, just top-line cost and speed.

Was this through a personal subscription? Last I looked into daybreak access, if required an organization based account.

> ...for a niche of customers, from whom the revenue is probably entirely negligible to OpenAI.

> Not saying it's right, ethical or anything like that, but it's not "stupid"

Quite many well-paying customers become very suspicious of whether integrating OpenAI into their workflows if they read about cases where OpenAI banned accounts willfully.

If OpenAI wants to become "infrastructure provider other companies should depend on (and thus pay)", OpenAI must behave like an infrastructure provider as for electricity or internet, i.e. must only ban accounts in exceptionally rare cases that 99.9 % of society will consider to be acceptable.

It creates a lot of uncertainty. Especially if you depend more and more on them in your day to day life.

If your always-on agent hears to wrong thing, you could lose access to things you depend on. The same reason makes GCP very unattractive.

They could have started with actually decrypting what was sent rather than a breathless 2 hours of monetised nonsense. Show me the http posts with the data that's actually sent.

A video on vulnerabilities in Asus routers was released yesterday: https://youtube.com/watch?v=kYnwCx2SNYY

But if you are talking TVs specifically, I expect they will get back to that too.

He bought a whole bunch of TVs and he’s testing them right now. Check out his YT channel.

Right......

Like none of these TV's hooks to an Open Wifi Hotspot and abuse you.

Man, it would really be interesting if they were caught using the speakers themselves as microphones.

Computer vision and categorization is ai. It isn't just generative AI.