Be firm; don't be afraid to use tough love. Spank it. Send it to boot camp. Three strikes you're out.
NEXT WEEK: Five warning signs your agent may be heading to wrong side of the tracks: bad grades, obsession with pop culture, delinquency, pornography, hallucinations...
What is the reason for the impairment that prevents anyone from being able to calculate a 10% tip by moving the decimal point one digit to the left? Or adding half of this for 15% or doubling it for 20%?
Ok too hard...
Doesn't every handheld already have calculator? Is there an impairment about this too?
Yet these impairments may not interfere with programming a custom tip calculator.
Or maybe the impairment prevents programming too, so fiddle with AI to conjur an app...
But then why not skip the fiddling and simply tell google to figure the tip?
At such point something strange should happen:
Won't even a the very unobservant notice the pattern of moving the decimal point one digit to the left?!
tldr This presentation is a puff piece that talks less about weight savings economy, and more about the motor's size-weight-to-power as comparable to high performance brakes and brags about application in fashionable super cars.
The conundrum of forgetting comes seems to come down to a gradient of persistence, whereby new stimulus prevails over memories.
If it didn't, the organism sensorium would be a captive to memories, which once noted seems like an obviously maladaptive dynamic.
Resolved: Perfect memory would preclude your sensory awareness...
So the next question is why does the gradient of persistence have such a shape / slope?
The dynamic of knowing vs. perceiving looks like some tradeoff over persistence and speed of access (a la caching).
Are memories something like deferred stimuli?
Thought must be some other domain within this gradient, subject to other dynamics?
What structures of the blob of protoplasm which is the brain manifest as Euclidian geometry, phonetic language / alphabet, and Turing machines? How does something so mushy give rise to such rigidity... which in turn bizarrely can model mushiness. Almost like an error condition.
AI boosters can help the world make better use of this tech by turning away come anthropomorphisms, such as the term "discover", which implies human agency, curiosity, and intellect, towards functionalisms, such as the term "uncover" which is a mechanical act.
It's a fact that AI is a tool and whatever it may do, it's under human guidance at the top level of control.
It's a common strong matter of opinion that were it possible to create sentient devices, it would be a crime to do so.
Anthropomorphisation is cartoonish, commonly borne of childish naiveté.
Such affectation is not appropriate in domains where life and livelihoods are at stake.
Computer security has been a nightmare since the advent of the Harvard architecture, due to stipulated indistinguishability of data and code.
AI architecture is extending this epic hazard towards infinity and beyond, where the construct of the prompt is hopelessly open-ended metadata over which security is meaningless, and agents, built on rusting perforated stacks, are prompting each other to do good knows what. No good is going to come of this.
So saying AI is "discovering" zero-days is cartoonishly tragic: AI is the right tool for the job; this work will never be done due to poor foundational designs.
One path back from the edge abyss is to drop the casual anthropomorphisms and at show a little more seriousness about chiding the path forward.
this seems like an entirely semantic disagreement you're having here. i don't see how it engages with the post's substantive claims. if it makes you feel better, imagine replacing "discover" with "uncover" = problem solved.
These kinds of pareidolic, co-dependent projections emitted by today's AI cognoscenti will be studied a hundred years from now for clues about how the advent of television skrod a republic starting with Nixon and ending in Trump.
The paper notes that 1899 was an extreme cold event -- an order of magnitude greater than average. Pick any other starting year and the trend becomes vastly smaller.
It's hard to tell if that's just bad data collection in 1899, perhaps a genuine cold snap exaggerated by data gatherers seeking extremes.
There's a similar magnitude of hot snap in 1936 that dwarfs every other one, and that's not simply cherry-picking on Christy's part, but it was confined to North America. Globally 1936 was not especially hot.
Regardless, as TFA says, "Extreme cold has decreased sharply." Cristy is trying to claim that extreme heat has also decreased, but I don't think he's evaluating the data critically.
Try inverting the argument: The current valuation is a continuation and acceleration of epic investment fraud.
To counter this point, someone has to show literal productivity gains from this tech that greatly exceed the valuation of Anthropic.
But what is productive about synthetically generated media?
If you are in the c-suite for today's most valuable companies, your entire enterprise valuation is based on confusion of media with production of tangible wealth. Moreover, you don't want this distinction to be sorted out because your valuation collapses. So you don't actually know what's real.
This's the deep, yawning, cavernous nature of the fraud.
BTW, whatever happened to NFTs.
When valuation comes out of nothing and manifests as nothing, it can puff away and not be missed, because it was never really there.
Upside is server chips may be completely free within a couple of years.
Looking at this demo, my gestalt is of answers to begged questions superimposed into a hyper Warner Bros Wile E. Coyote cartoon.
What it demonstrates is that genAI is getting ever more powerful at emitting stuff that looks like other stuff, but the uncanny valley is deepening with each step, where now the Coyote has run off the edge of the void cliff, but isn't yet falling because he hasn't yet realized he's run off the edge.
On this forum, I routinely see the question "So what is AI doing that's real work"? The answer can't be "passing the otter test...", can it?!
When the writeup claims "I would have been very happy if this paper was the outcome of a 2nd year PhD project" my heart sinks and how meaningless doctoral work must be to say this without irony.
By the way, that harbor town model is not "beautiful" unless you think that any image with a color scheme of complementary gradients is beautiful.
As to the observation that this appliance will be great for making Powerpoint presentations...
(Zzzz --Huh? Oh is the presentation over? let's go to the pub!)
Parenting the troubled agent:
Be firm; don't be afraid to use tough love. Spank it. Send it to boot camp. Three strikes you're out.
NEXT WEEK: Five warning signs your agent may be heading to wrong side of the tracks: bad grades, obsession with pop culture, delinquency, pornography, hallucinations...
reply