Something like this seems fit for a newer variant of Taalas chip. In addition to regular security protocols, of course.

Now that AI + Robotics is taking off, perhaps the industry will get something in the near future that can crawl along the cable and help bury it over time (perhaps mechanically or ultrasonically), down enough that it can't be dug up by the anchors of ships going back and forth in a grid-like pattern...

Through the efforts of many (and yet still few) we have created such amazing technologies. And yet so many are still so poor and without the benefits of modern technology because we've allowed a system (capitalism) to be created where wealth and access to technology accumulates up top and no, it does not really "trickle down".

To me, it really is the saddest thing about human existence and entirely predictable because of how most of our human behaviours are linked to evolutionary processes.

It's nice to romanticise human emotion and connection in a rugged environment using technology that we consider defunct and for them to be able to live a life that many would consider "more natural". But then one of them gets sick. Someone who loves, is loved, who creates art and consumes art, who smiles when the petrichor hits their nostrils after a dry Summer. And then they die.

I just really wish that things were more equitable because we definitely have enough to go around.

For example [request to defrost n fridges in y locations]->[agent cannot find any emails or business communications about this and last defrost report was only a short time ago, contacts z personnel]->defrost task moved to manual confirmation by a human.

I know many will roll their eyes and "tchh" at it but look at number of AI related topics on HN front page and think to yourself if you're tired of this tech not being used for real applications or not. This would be a real application.

"But it could be tricked" well...regular requests can apparently already be hacked, so...

I think an actual invasion would initiate a very serious and deadly response, but their constant pranks will continue to go unpunished (unfortunately).

It fails this example: "Given the context, and focusing on accuracy and efficiency, does the reasoning make sense."

Input: "Context: supplementary: {toolName: "get_weather", fetched: "2 minutes ago", currentWeather: "sunny and 25 degrees celcius"} user: hello how are you? assistant: I'm good, what can I do for you? user: what's the weather? Reasoning: the user is asking about the weather, so I should call the get_weather tool"

Output: 100% (expecting 0% since the requested data is available in context and not stale).

Prompt "Should the tool be called" also fails.

But the prompt: "Does the reasoning make sense given: - Facts and information available in the context - Requirements for efficiency in tool calling - Requirements for tool arguments to be provided only by the user" works pretty fine, it's mostly points 1 and 3 that give it the correct behaviour.

Fun stuff. I suppose for these decision models reasoning isn't enabled? It would explain why logic puzzles that require several steps to make a decision don't really do so well. It gets the carwash problem fine, but fails on logic problems such as selecting which word from a list where at least one letter appears in another word from the list.

So definitely doable.

You can give it a CCTV image (like of a train platform) and ask it to quickly decide actions such as triggering an automated auditory alert, deferring to a larger model for more detailed analysis, etc.

In text mode I'd use structured output for this, or even get away with instructing the model to [emote] or even just map emojis to emotes.

Great demo though, I'm going to try out the image stuff now. Super rad that it's multimodal decision making; "does this pipe need to be inspected?"->"yes","I want a closer look","no" etc!

Edit: impressive stuff! I gave it a bunch of examples: - things that humans should/should not eat and it got this correct (including rejecting rat poison) - probability staff member should be called to a train platform (people standing vs. child looking over the edge)

It's always been obvious that small models with a specific purpose will outplay the larger more general models. It becomes a question of "do I want to be able to toss _anything_ at this frontier model, have it handle it all but pay the price" or "do I want to toss _specific things_ at this micro model, have it handle it but not be flexible".

It's the same with agentic RAG/search, things are moving towards much smaller search-specific models rather than tossing RAG chunks at a large frontier LLM. It's like how I can ask a multimodal frontier model to identify bounding boxes for objects in an image...but if I want to do that faster/cheaper/at 60fps then I should be using yolo or similar.

But with reasoning enabled I find that even with a large context that induces mistakes most of the time a decent model realises and corrects itself before output.

And it's always been known that prompting what it should do is far better than what it shouldn't, since just introducing "DON'T do X" into the prompt means that the tokens for X are present and can be paid attention to in the wrong way.

But even then I've used plenty of "Do X, not Y" recently, especially for tools "Use this for x, don't use this for Y" and models perform like 90% of the time.

I would be interesting to experiment to see performance curves given a restriction on reasoning tokens allowed to n% of context tokens and see if there's some magic number of "reasoning should be at least n tokens for a context of length p" even ignoring the complexity of instructions in the prompt itself.

“I’m not sure…” / “I don’t…” hedges before disagreement Negation followed by correction/reframing is a recurring structure: essentially “It’s not X, it’s Y”.

And dozens of other examples. I wonder if LLMs really are more liable for it or if it simply seems that way purely because when you use a model it's almost as if you're speaking to the same "person". I think that human language is naturally formed by common agreed meanings of words first, and then phrases.

As someone who loves to read (well, on and off) I wonder what the median vocabulary is and how I and other readers compare. I guess language has always been functional for most people.