https://thephilosophicalsalon.com/the-end-of-algorithmic-denialism/
It seems Professor Bender, et. al.'s work has gotten under some folks' skin.
>When it comes to denying cognitive value, the authors declare these tools “inherently unreliable, being designed to make shit up”. When it comes to documenting harm, they implicitly presuppose that the models work well enough to be adopted at industrial scale, to take commissions away from professionals, to pollute the information ecosystem. The two claims require incompatible premises; the book alternates between them according to rhetorical convenience without ever acknowledging the contradiction.
Those two claims are concordant with each other. LLMs are inherently unreliable. The commissioners of its products don't care. The resulting harm is real. There is no incompatibility or contradiction here.
>If, on the other hand, we admit that at least some users are capable of distinguishing useful output from useless output, and that they do so regularly by selecting, discarding, and correcting, then the usefulness of the models is an empirical fact and the thesis that they “do not work” collapses.
The usefulness of the *models*? The *models* aren't doing the distinguishing. The *user* is. The fact that the distinguishing needs to happen is the evidence of the model's unreliability.
>Crawford, like Bender, applies to AI a moral standard that she does not apply to other technologies. Nearly all of our technologies rest on analogous systems of extraction, consumption, and pollution, often on an enormously larger scale; yet we rarely subject intercontinental flight, video streaming, or 3D rendering to the same critical rigour.
The externalities of the tools du jour matter. Intercontinental flight is routinely subjected to critique of its externalities. Video streaming and 3D rendering both require fewer computational resources (I'd wager at least an order of magnitude fewer) than the training of AI models. That every other consumer of integrated circuits is stuck waiting in line to pay more for the same product, while AI data centers hoard the very foundations of general purpose computing itself (CPUs, GPUs, memory, storage), should be sufficient externality for observers to question just how useful the AI tools actually are in comparison to the known values of existing tools and workflows.
>The conclusion of The AI Con reveals the structural limit of the entire approach. Bender and Hanna propose “strategic refusal” as a political horizon: saying no to AI, drawing inspiration from Luddite movements and feminist struggles. The invitation is suggestive, but the question left unanswered is: refusal in favour of what? If AI is a technology that, as the book itself documents, is already reshaping work, healthcare, education, and cultural production, refusal without an alternative project risks being a symbolic gesture that leaves the field open to those who govern and deploy these technologies.
In favour of what we know works: what we were all doing less than a decade ago, before hyperscalers were in search of their next big growth opportunity. That is, human beings producing works with far greater reliability than AI can. We need not indulge the ruling class when it wishes to squeeze ever more out of us.
>The position I continue to hold is that these tools have concrete potential, and that precisely for this reason the politically relevant battle concerns who controls them and how they are governed. The denialism of the early period has delayed this discussion, sustained by the illusion that we are dealing with nothing more than an ephemeral and overrated fad. It is time to abandon what Benjamin Bratton has called the first stage of grief towards AI: denial.
The financials of AI alone should caution against a belief of this being a genuinely new stage of computing that's not only here to stay but here to become the new normal. Zitron's exhaustive work on just how out-of-wack AI's finances are makes a strong case that AI is a bubble.
CC nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpqe0a6n84krygh2qc3ajptu4fa4ma0qy53acz8smsyhztgxljxkghsk25876 (nprofile…5876)
