AI doesn’t actually work

New / Sales espionage
May 8 11 Comments

Does anyone else work for an AI company where the claims far outstretch the reality of what your tech does. I haven’t seen AI at least in sales tech actually work

comments

Want to comment? LOG IN or SIGN UP
TOP 11 Comments
  • Veritas thetruthin
    It does work, but most AI companies are not really AI companies, they just repackage stuff, with obviously modest results.

    However look at Google, Facebook, Uber, Apple. They have been extremely successful at deploying AI at scale.
    May 8 1
    • Walmart.com rand.randm
      Uber and Apple are no where close to where G FB and MSR are. Apple recently recruited Ian Good fellow and had recruited John G and a prof. From CMU.
      May 10
  • Adobe HaroldWren
    @espionage you are in sales, it is your kind who overhype this. Those of us who do “AI” call it “machine learning” and ship it with a long list of limitations. (This is reductionist of course, but those who actually work with “AI” would tell you it works exactly as well as we think it does, those who sell... well ...)
    May 8 1
    • New / Data
      NullsBohr

      New Data

      PRE
      Mu Sigma Inc.
      NullsBohrmore
      Big facts 💯 right here !
      May 8
  • Google -Knight~
    It works and in wonders . There be iterations to it though
    May 8 0
  • New NcEf26
    Most companies are bad at it, and they hire based off pedigree for data scientist roles. So they get a PhD in Physics who has no real skills, and let him loose. The results aren’t much different than a python notebook pulled from GitHub.
    May 9 0
  • Microsoft Gorillala
    Define AI because I see more statistical math than intelligence in these companies.
    May 8 0
  • Walmart.com rand.randm
    It works for companies with real AI talent and data. There are just a few that can afford both :-). For most startups it is just a hype especially the ones with no ML talent in executive leadership. I have spent hours explaining to ambitious founders and investors that in applied ML data is the king. Rarely a startup can create ground breaking tech that can compete with dedicated Research talent at MSR, G Brain or FAIR.
    May 10 0
  • Uber / Eng ͡° ͜ʖ ͡°
    Mechanical Turk was ahead of its time.
    May 8 0
  • Nvidia yesm
    It works only for companies with huge amounts of data. So yeah - Google Facebook Microsoft etc
    May 8 0
  • Google GCPCE
    My old startup massively oversold what it could do with AI, and most enterprise companies have done a lot of PoCs with ML, but very few have put things in production. It can and does provide value though in well thought out use cases.
    May 8 0