Meta dropped three new AI fashions over the weekend: Scout, Maverick, and the still-training Behemoth, billed as the following evolution of “open-ish” AI. However as an alternative of pleasure, the response was largely shrugs. Critics known as the discharge underwhelming, saying it lacked the sting anticipated in at this time’s breakneck AI race. Meta’s clear try to claw again some consideration rapidly turned messy. Accusations started circulating on X and Reddit round benchmark tampering, a thriller ex-employee, and huge gaps between the fashions’ private and non-private efficiency.
At this time, on TechCrunch’s Fairness podcast, hosts Kirsten Korosec, Max Zeff, and Anthony Ha are unpacking Meta’s rocky rollout, the AI trade’s obsession with wanting sensible on paper, and why, as Kirsten put it, “creating one thing to do effectively on a take a look at doesn’t at all times translate to good enterprise.”
Hearken to the total episode for:
Fairness can be again subsequent week, so keep tuned!
Fairness is TechCrunch’s flagship podcast, produced by Theresa Loconsolo, and posts each Wednesday and Friday. Subscribe to us on Apple Podcasts, Overcast, Spotify and all of the casts. You can also comply with Fairness on X and Threads, at @EquityPod. For the total episode transcript, for individuals who desire studying over listening, try our full archive of episodes right here.
