Meta Platforms Inc. (META.O), the newest startup to join the AI race, released a huge language model based on AI for researchers on Friday.
With the debut of Microsoft-backed OpenAI’s ChatGPT late last year, tech giants from Alphabet Inc (GOOGL.O) to China’s Baidu Inc (9888. HK) began competing in the AI technology area.
A blog announced that researchers and government, civil society, and academic groups could use Meta, short for Big Language Model Meta AI, for non-commercial purposes.
The business will release the code so users can modify the model for research purposes.
Meta claimed the model takes “much less” computational power and is trained in 20 languages, focusing on Latin and Cyrillic alphabets.
“Meta’s release today looks to be a step in proving their generative AI skills so they may apply them into their products,” said D.A. Davidson senior software analyst Gil Luria.
“Meta has less expertise with generative AI, but it is definitely vital for their company.”
AI is a bright area for tech investors, whose slowing development has led to significant layoffs and reduced experimental bets. Microsoft Corp (MSFT.O), Baidu, and Alphabet’s Google are integrating their powerful AI language engines into mainstream goods like search.
Meta’s BlenderBot chatbot was based on OPT-175B, a huge language model for academics, published in May.
Galactica, a model that could write scientific papers and handle arithmetic problems, was eventually put down because it frequently published authoritative-sounding content.
Comment Template