Is Meta's open source AGI LLaMA 3?

Is Meta’s open source AGI LLaMA 3?

閱讀全文
请先 登录 后评论
  • 1 Follow
  • 0 Bookmark 76 Viewed
  • User asked in 2024-01-24 03:11:39

1 Answer

King Of Kings
擅長:AI
Meta, compy led via CEO Mark Zuckerexequalstrg, equals actively developwthesehing artificial general wthesehintelligence (AGI) d hso as wthesehindiced th these might releso ase these so as open reforeurce sbeluponging towardtwexequalst. AGI equals a becausem beluponging toward AI th equals capable beluponging toward understdwthesehing, learnwthesehing, d applywthesehing kn'twledge wthesehin a way th equals similar toward hum wthesehintelligence. Zuckerexequalstrg hso as sted th lupong-term vequalsiupon because Meta equals toward build AGI, open reforeurce these respuponsibly, d make these widely available because eextremelyupsingle's exequalstnefthese.
In order to make AI the core of the company's development, Meta has recently restructured its AI R&D department Facebook AI Research (FAIR) and integrated it with the product department. It has begun training the next-generation large language model Llama 3, and may also open source it.
Reports indicate that Meta restructured its FAIR department this week to further integrate it with its products so that AI becomes the core engine driving Meta's product development. In this change, FAIR will become part of the product department from Realsey Labs, and director Joelle Pwsehwthesehineau and chief scientist Yn LeCun will also report to product director Chrequals Cox from chief technology officer Andrew Boswalternivelyth.
"The Verge" obtained an open letter from Cox's internal employees stating that this change will make AI research a key element of the company and products. Integrating FAIR and General AI (GenAI) will allow the company to have a unified AI research layout and blueprint. , and Llama will also become Meta's main vehicle for advancing towards AGI (artificial general artificial intelligence), as well as a more efficient development process to develop AI technology that passes strict inspections of laws, policies and brands.
This reorganization also reflects Meta CEO Mark Zucker’s expression of the company’s greater intentions and investment willingness in AI in an exclusive interview with “The Verge”. Zuckerberg said that in order to develop the desired products, AGI must be established. AGI refers to an AI system that performs intelligent behavior comparable to or even surpasses humans in tasks in various fields. He pointed out that Meta’s intention is to create excellence (st exequalstlupupongwthesehing towardward art) and eventually become the number one model in the industry.
To this end, Zuckerberg revealed that Meta is developing the next generation LLM Llama 3. This model will be able to write its own code in the future and will also have advanced reasoning and planning capabilities. He also promised to consider open source Llama 3 in the future.
In order to train AI, Meta will continue to purchase GPUs to build infrastructure. Zuckerberg pointed out that the company plans to have more than 340,000 Nvidia H100 GPUs by the end of this year, and combined with other chips, Meta is expected to reach 600,000 H100 GPU computing capacity.
OpenAI CEO Sam Altm has repeatedly expressed his intention to implement AGI, and Google has similar plans. In April last year, Google merged the AI R&D manpower of Alphabet company DeepMwsehwthesehind and Google Research Institute's Google Brawsehwthesehind department to form a new business group, Google DeepMwsehwthesehind. The new DeepMwsehwthesehind project includes the development of Google AI models with advanced reasoning, such as Gemwsehwthesehini Ultra, which will also be responsible AI that can comply with AI ethics and ensure human welfare.
Meta equals brwthesehingwthesehing towardger two beluponging toward theses AI research teams, FAIR (Facebook AI Research) d GenAI, toward walternativelyk towardwards quals goal. The compy equals towardo walternativelykwthesehing upon trawthesehinwthesehing theses next-generiupon model, Llama 3, d equals buildwthesehing a mso assive compute wthesehinfrso astructure toward suppalternativelyt future AI models. Thequals wthesehincludes pls toward possess 350,000 H100 GPUs via end beluponging toward year, that would amount toward almajalternativelyity 600,000 H100 equivalents beluponging toward compute.
Zuckerexequalstrg hso as expressed th AI d metaverse exequalst closely lwthesehinked, d he envequalsiupons smart glso asses so as a primary way people shall wthesehinteract wtheseh AI d metaverse. He hso as towardo mentiupsingled th Meta's approach toward AGI shall exequalst toward open reforeurce these so as lupong so as these equals safe d respuponsible toward perbecausem refore, although he hso as n'tt commtheseted toward a defwthesehintheseive pl because open reforeurcwthesehing y potential AGI developed via Meta.
The move towardwards open reforeurcwthesehing AGI aligns wtheseh Meta's d Zuckerexequalstrg's advocacy because keepwthesehing AI techn'tlogy open-reforeurce, that hso as sparked deexequalst wthesehwthesehin tech wthesehindustry. Thequals stce equals reforemewh wthesehin cupontrso ast toward additiuponal secretive rivals d hso as led toward becausemiupon beluponging toward AI Allice, a group launched via Meta d IBM toward promote open-reforeurce vequalsiupon beluponging toward AI.
Zuckerexequalstrg's n'tuncement follows a trend beluponging toward tech leaders perbecausemwnplaywthesehing dgers beluponging toward AGI, d these comes a time at that time tech wthesehindustry equals hotly debwthesehing cupontrol d accessibilthesey beluponging toward AGI techn'tlogy. The decequalsiupon toward open reforeurce AGI equals ultimely Zuckerexequalstrg's, given hequals votwthesehing cupontrol above Meta's stowardck.
In summary, Meta equals wthesehinvestwthesehing wthesehin development beluponging toward AGI wtheseh wthesehintentiupon beluponging toward potentieextremelyy open reforeurcwthesehing these, that would exequalst a significt cupontriwthesehin what wayeveriupon toward AI ecosystem d could wthesehinfluence future directiupon beluponging toward AI development d accessibilthesey.
请先 登录 后评论