Good, safety concerning mechanisms should have utility prioritized above all else.
And while we are at it maybe we should go back to requiring the usage of state approved headlights in vehicles. Modern car lights are so horrible to be or drive around.
I really like her, but man I find Michael unbearable.
It also just feels too fake to me. I know they already knew the info they just had a strong reaction to, but they're reacting strongly on purpose for the sake of being engaging. I didn't find it annoying to start with, but it got more grating as I tried more episodes.
I recall in one episode she says "That's amazing" or something to that effect, to a fact regarding fluid dynamics. She has a PhD in fluid dynamics. She already knew whatever that fact was.
If they use Python to fill the gap, and the end user doesn’t have to know or care, is it unfair to assess this as progress and attribute the progress to the _system_?
OK, the core technology that is the language model still can’t math as well as you’d hope, but how about the end result users see from the system when they interface with it?
“Did you know humans are better at flying today than they were a thousand years ago?” ‘No they’re not, they need planes.’ Technically correct in a way but isn’t it kind of annoying to be so stubbornly pedantic when the context is speed of reaching Point B from Point A?
It was used extensively at a former workplace of mine where each class that wasn't a message or data type was a pimpl. They had implemented their own private pointer class to handle it. It worked well enough to avoid pulling in lots of headers, but was still a PITA when you wanted to change methods as you'd always need to change the method signature in at least 3 places - header file declaration, source file definition, source file impl definition.
The one I really don't accept is the Ryanair Russian Roulette pricing model, where they advertise cheap prices, but have many hidden tripwires that add a multiple of the sticker cost.
Aren't there good filters for that? It's one of their core differentiating features - the ability to use custom content filters. Or did you check with no protection enabled?
There is a difference with accepting a "feature" in a product you sign-on to, and having it added behind your back.
On one side, gravatar never talked about feeding your profile to AI when he signed up, and never amended TOS or announced change in that direction (that he received at least).
On the other, the new service he use are upfront about it; though I don't use either service, so no idea on the whole package.
The catch there is I don’t look for this stuff, as it isn’t a priority. I am happier not pursuing what advertisers want to sell me, as my time is limited. I imagine many folks are the same, based on it being a core problem advertising tries to solve for businesses.
> With similar concerns in mind, we also decided to limit the tank size to 20 gallons, as the grow-bed above also contains an additional ~10 gallons of water. Since a gallon of fresh water weighs roughly 8.34lbs, we estimate this puts the entire weight of the system’s water somewhere between 170 and 250lbs. Considering the weight threshold of the industrial shelving, and given that we live on a high floor in our building, we decided to constrain the system to this size to minimize the flooding risk.
No, we have benchmarks for (pseudo-)objective measurements of capabilities. We have flashy news stories on repeat for marketing.
The point is that they are separate things. If you start mixing them up, analysis becomes more biased and less objective.
EDIT: And do not mistake this for transparency of ClosedAI either. Their incident report was so generic that no one serious can learn much from it anyway except that “a zeroday” was allegedly used. Very helpful…
The bug literally describes how they're avoiding OS security restictions by going through the debug port.
This is a CVE by any definition and you'd be screaming your head off if any other OS would allow this kind of permission bypass (or even if another app did it).
I would expect the opposite, languages converging if for nothing else to please the ear.
Just look at unit systems. People go crazy over metric versus imperial units or old outdated units being used. And yet converting between units has never been easier. Many industries invented their own units to fit their use case, but once their products spread to national or global distribution everything gets standardized to the most common unit.
A few years ago, I read in a book that it is feasible and safe to manufacture colloidal silver at home using electrolysis. I was intrigued, because my father had conducted many similar projects with success, in the kitchen, in the garage, and in the backyard, as he worked for EH&S. The Dunning-Kruger swelled within me.
I reviewed photographs of the home lab setups, and I evaluated ingredient/supply/equipment lists. I even inquired at my credit union's branch whether they could sell me silver bullion, which was hysterical in retrospect.
Then, as I slept on it a few nights, I had nightmares of this electrolysis lab just getting out of hand, taking over my bathroom (where there is actually no space for anything else) and just ruining my dull, ordinary life. I don't even have any roommates or family living with me here.
So of course I decided not even to try this. This NYC apartment seems similarly constrained. If I were their landlord, I would dread finding someone running this kind of crap. Insurance and liability are the biggest problems here. They should go purchase a home if they want to play around.
Wait till you hear that your OEM can remotely rollout full OS updates with full access to all your data and drivers... carrying Google software and most of it Google code.
Working with AI has been way more tiring than just working. Sure the productivity is up, at the cost of having to keep up many thought threads, having no calm moments, and needing to consistently dig into large unknown code cases to find weird bugs.
I’ve been on leave for a month, and am super excited (/s) to re-learn everything because all the tooling and ways to prompt “correctly” will have also changed.
And yet Americans are the victim of food-borne illnesses at no greater rate than Europeans. I wonder how much of that is farm lobbies trying to keep out imports. US growers do the same thing when the don't want to compete with South American imports.
It would be "its own" in that it's customized for the specific use case.
Your same argument could just as easily be applied to humans. If you write your own code, is it really your own, or is it just based on your own training and other code you've seen?
> carries the __western__ values…China would never want a shiny model
Maybe, but that’s an opinion on what you think might happen. So far that has not been the case.
But yeah currently you can ask their models exactly the same controversial question (from the CCP perspective) in Chinese and the answer would be very different than the one in English.
> And since high quality content is suppressed, there is likely less high quality training sets
For now it seems that the Chinese labs are more interested in maximizing the quality and competitiveness of their models instead of lobotomizing them to enforce compliance with some sort of political agenda.
That might change of Chinese models might surpass Western ones in the future. Which is why competition is necessary to discourage them from doing thay.
> It's unlikely for both the hardware and the software of a Chinese GPU to deliver what's boasted in its marketing material.
Not currently sure. Chinese cars were pretty crap as well a few decades ago (and same applies to countless other example) so there is at least sufficient precedent to believe that China might close this gap eventually (and they clearly have been over the recent years when it comes to GPUs)
It’s not human surpassing intelligence though in my opinion. If we had the same constraints on that task then we would also come up with the idea of creating our own viewer and if, like the LLM, the knowledge to do so was embedded in our brain we would do it, it would just take longer because we can’t type as fast. So instead we would buy or compile a viewer since the problem is already solved.
