News for hackers. Stuff that matters.
Re:Google scrambles to manually remove weird AI answers in s... by JCM9 on Saturday May 25, 2024 @12:10PM (#40475874)
The fall of Google’s reputation on ML is nothing short of spectacular. They went from having a near untouchable reputation as being far ahead of any other large tech company on ML to total shambles in a year. Everything they’ve released has been a complete popcorn worthy dumpster fire from faked demos, to racist models that try and pretend white people don’t exist, to this latest nonsense telling me put glue on my pizza.

What the heck happened? Or was their reputation always just more hype than substance?

Re:Google scrambles to manually remove weird AI answers in s... by rvnx on Saturday May 25, 2024 @12:15PM (#40475906)
It could be because they actually released something. If you look back, the Google Research blog posts always have grandiose claims, but you can often never use them.
Re:Google scrambles to manually remove weird AI answers in s... by CydeWeys on Saturday May 25, 2024 @12:20PM (#40475953)
AlphaGo, AlphaFold, and Waymo FSD are all released in the sense that you can see them actually working in the real world. Those all took much longer to put together than whatever rushed features were released to catch up with OpenAI, however.
Re:Google scrambles to manually remove weird AI answers in s... by tadfisher on Saturday May 25, 2024 @02:33PM (#40476924)
They are also extremely constrained problem spaces relative to the problem space of LLMs, which is apparently "everything imaginable".
Re:Google scrambles to manually remove weird AI answers in s... by padthai on Saturday May 25, 2024 @05:08PM (#40477891)
Waymo is not Google. And Deepmind operated quite independently until not long ago.
Re:Google scrambles to manually remove weird AI answers in s... by ugjka on Saturday May 25, 2024 @12:17PM (#40475934)
research != product
Re:Google scrambles to manually remove weird AI answers in s... by bbarnett on Saturday May 25, 2024 @12:31PM (#40476046)
At least Elmer's white glue is edible, millions of kids agree.

(The logic sort of makes sense. Glue sticks things together, and some glue is edible.)

Re:Google scrambles to manually remove weird AI answers in s... by seydor on Saturday May 25, 2024 @12:15PM (#40475911)
It's not really that bad. I use gemini often and it's great. I prefer their UI
Re:Google scrambles to manually remove weird AI answers in s... by tymscar on Saturday May 25, 2024 @01:29PM (#40476467)
What do you like more about their ui?
Re:Google scrambles to manually remove weird AI answers in s... by seydor on Saturday May 25, 2024 @01:31PM (#40476481)
faster, it has options like 'modify'. I also feel it follows my commands better, esp. when i ask to rephrase
Re:Google scrambles to manually remove weird AI answers in s... by arccy on Saturday May 25, 2024 @12:19PM (#40475948)
research != product
Re:Google scrambles to manually remove weird AI answers in s... by calebkaiser on Saturday May 25, 2024 @12:40PM (#40476122)
There was an interesting interview with David Luan about this recently. For context, he was a co-lead at Google Brain, early hire at OpenAI, and is now a founder at Adept: https://www.latent.space/p/adept

The TL;DR on his take is that there are organizational and cultural issues that prevent Google from focusing their research efforts in the way that is necessary for what he calls "big swings," like training GPT-3.

In regards to your second question, Google's reputation in ML is definitely not hype. Purely on the research side, Google has been behind some of the most important papers in modern ML, particularly around language model. The original Transformers paper, BERT, lots of work around neural machine translation, all of the work that DeepMind has done post-acquisition, and the list goes on. On the applied side, they also have some of the most successful/widely-adopted ML-powered products on the market (think RankBrain/anything involving a recommendation engine, Translate, Maps, a ton of functionality in Gmail, etc).