The issue: Big Tech has the $$ infrastructure, data, ability to pay talent, and access to market which no one else does. So as academics, you either pay retail for access, or get it discounted/free by yoking yourself to Big Tech (via dual affiliation, or just being hired). 2/
MARKET TRENDS
-
Seven GPT-4 Competitors, Price Wars, Modest AI Adoption
By
–
Prediction: By end of 2024 we will see
• 7-10 GPT-4 level models
• No massive advance (no GPT-5, or disappointing GPT-5)
• Price wars
• Very little moat for anyone
• No robust solution to hallucinations
• Modest lasting corporate adoption
• Modest profits, split 7-10 ways -
Two GPT-4 Level LLMs Released in One Week
By
–
Two GPT-4 level LLMs in a week… let's see how OpenAI responds
-
GenAI dominates 80% of AI investment, stifles alternatives
By
–
GenAI is probably 80%+ of AI investment right now. Maybe more. That is the monopoly I was and am referring to. Some of it is from VC, some internal from companies. Very little oxygen for alternative approaches right now.
-
Monopoly on AI Investment Sector Discussed
By
–
It’s a monopoly in a sector, viz on *investment in AI* I didn’t say people didn’t (or wouldn’t in my alternative history) still build tractors or fintech or whatever. Very uncharitable reading, put in rather kind words.
-
Ladder Builders Cult Blocks Rocket Innovation Progress
By
–
Did anyone ever argue that ladders aren’t useful? Gonna write an alternative history in which we never make it to the moon because of a cult of ladder builders monopolize venture capital and public imagination and shout down rockets as implausible, expensive, and unproven.
-
Driverless Cars Remain Demos, Not Production Ready
By
–
yes i think you skimmed it. and honestly if you think driverless cars are anything more than a demo you aren’t paying attention.
-
AI Company Faces Competition, Legal Troubles, Strategic Challenges
By
–
Sure but they also have already 4 close competitors that have largely caught up, with at least one competing on price, a dozen or so lawsuits, upcoming licensing costs, at least 2 or 3 federal investigations, no moat, high operating costs, and a desperate need to build something
-
Claude 3 family outperforms GPT-4 on key benchmarks
By
–
1/ Claude 3 Consists of a family of three models (Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus); Claude 3 Opus (the strongest model) seems to outperform GPT-4 on common benchmarks like MMLU and HumanEval.