OpenAI

01OpenAI is an American Artificial Intelligence, AI, research organization founded in December 2015 and headquartered in San Francisco, California.

02Its stated mission is to develop safe and beneficial artificial general intelligence, AGI, which it defines as highly autonomous systems that outperform humans at most economically valuable work.

03As a leading organization in the ongoing AI boom, OpenAI is known for the GPT family of large language models, the DALI series of text-to-image models, and a text-to-video model named Sora.

04Its release of ChatGPT in November 2022 has been credited with catalyzing widespread interest in generative AI.

05The organization consists of the non-profit OpenAI Incorporated, registered in Delaware, and its for-profit subsidiary introduced in 2019, OpenAI Global, LLC.

06Microsoft owns roughly 49% of OpenAI's equity, having invested 13 billion US dollars.

07It also provides computing resources to OpenAI through its cloud platform, Microsoft Azure.

08In 2023 and 2024, OpenAI faced multiple lawsuits for alleged copyright infringement against authors and media companies whose work was used to train some of OpenAI's products.

09In November 2023, OpenAI's board removed Sam Altman as CEO, citing a lack of confidence in him, but reinstated him five days later after negotiations resulting in a reconstructed board.

10Many AI safety researchers left OpenAI in 2024.

11History 2015-2018 Non-Profit Beginnings In December 2015, OpenAI was founded by Sam Altman, Elon Musk, Ilya Sutzkever, Greg Brockman, Trevor Blackwell, Vicky Chung, Andre Karpathy, Dirk Kingma, John Shulman, Pamela Vagata, and Wycheck Zaremba, with Sam Altman and Elon Musk as the co-chairs.

12A total of $1 billion in capital was pledged by Sam Altman, Greg Brockman, Elon Musk, Reid Hoffman, Jessica Livingston, Peter Thiel, Amazon Web Services, AWS, Infosys, and YC Research.

13The actual collected total amount of contributions was only $130 million until 2019.

14According to an investigation led by TechCrunch, Musk was its largest donor while YC Research did not contribute anything at all.

15The organization stated it would freely collaborate with other institutions and researchers by making its patents and research open to the public.

16OpenAI was initially run from Brockman's living room.

17It was later headquartered at the Pioneer Building in the Mission District, San Francisco.

18According to Wired, Brockman met with Yoshua Benjo, one of the founding fathers of deep learning, and drew up a list of the best researchers in the field.

19Brockman was able to hire nine of them as the first employees in December 2015.

20In 2016, OpenAI paid corporate-level, rather than non-profit-level, salaries, but did not pay AI researchers' salaries comparable to those of Facebook or Google.

21Microsoft's Peter Lee stated that the cost of a top AI researcher exceeds the cost of a top NFL quarterback prospect.

22OpenAI's potential and mission drew these researchers to the firm.

23A Google employee said he was willing to leave Google for OpenAI partly because of the very strong group of people and, to a very large extent, because of its mission.

24Brockman stated that the best thing that I could imagine doing was moving humanity closer to building real AI in a safe way.

25OpenAI co-founder Wycheck Zaremba stated that he turned down borderline crazy offers of two to three times his market value to join OpenAI instead.

26In April 2016, OpenAI released a public beta of OpenAI Gym, its platform for reinforcement learning research.

27NVIDIA gifted its first DGX One supercomputer to OpenAI in August 2016 to help it train larger and more complex AI models with the capability of reducing processing time from six days to two hours.

28In December 2016, OpenAI released Universe, a software platform for measuring and training an AI's general intelligence across the world's world's supply of games, websites, and other applications.

29In 2017, OpenAI spent $7.9 million, or a quarter of its functional expenses, on cloud computing alone.

30In comparison, DeepMind's total expenses in 2017 were $442 million.

31In the summer of 2018, simply training OpenAI's Dota 2 bots required renting 128,000 CPUs and 256 GPUs from Google for multiple weeks.

32In 2018, Musk resigned from his board of directors seat, citing a potential future conflict with his role as CEO of Tesla due to Tesla's AI development for self-driving cars.

33Sam Altman claims that Musk believed that OpenAI had fallen behind other players like Google, and Musk proposed instead to take over OpenAI himself, which the board rejected.

34Musk subsequently left OpenAI.

35In February 2019, GPT-2 was announced, which gained attention for its ability to generate human-like text.

362019, transition from non-profit.

37In 2019, OpenAI transitioned from non-profit to capped for profit, with the profit being capped at 100 times any investment.

38According to OpenAI, the capped profit model allows OpenAI Global LLC to legally attract investment from venture funds and, in addition, to grant employee stakes in the company.

39Many top researchers work for Google Brain, DeepMind, or Facebook, which offer stock options that a non-profit would be unable to.

40Before the transition, public disclosure of the compensation of top employees at OpenAI was legally required.

41The company then distributed equity to its employees and partnered with Microsoft, announcing an investment package of $1 billion into the company.

42Since then, OpenAI systems have run on an Azure-based supercomputing platform from Microsoft.

43OpenAI Global LLC then announced its intention to commercially license its technologies.

44It planned to spend the $1 billion within five years, and possibly much faster.

45Altman has stated that even a billion dollars may turn out to be insufficient, and that the lab may ultimately need more capital than any non-profit has ever raised to achieve artificial general intelligence.

46The transition from a non-profit to a capped profit company was viewed with skepticism by Orin Etzioni of the non-profit Allen Institute for AI, who agreed that wooing top researchers to a non-profit is difficult, but stated I disagree with the notion that a non-profit can't compete and pointed to successful low-budget projects by OpenAI and others.

47If bigger and better funded was always better, then IBM would still be number one.

48The non-profit OpenAI Inc.

49is the sole controlling shareholder of OpenAI Global LLC, which, despite being a for-profit company, retains a formal fiduciary responsibility to OpenAI, Inc.'s non-profit charter.

50A majority of OpenAI, Inc.'s board is barred from having financial stakes in OpenAI Global LLC.

51In addition, minority members with a stake in OpenAI Global LLC are barred from certain votes due to conflict of interest.

52Some researchers have argued that OpenAI Global LLC's switch to for-profit status is inconsistent with OpenAI's claims to be democratizing AI.

53In 2020, OpenAI announced GPT-3, a language model trained on large internet datasets.

54GPT-3 is aimed at natural language answering questions, but it can also translate between languages and coherently generate improvised text.

55It also announced that an associated API, named simply the API, would form the heart of its first commercial product.

56Eleven employees left OpenAI, mostly between December 2020 and January 2021, in order to establish Anthropic.

57In 2021, OpenAI introduced DALI, a specialized deep-learning model adept at generating complex digital images from textual descriptions, utilizing a variant of the GPT-3 architecture.

58In December 2022, OpenAI received widespread media coverage after launching a free preview of ChatGPT, its new AI chatbot based on GPT-3.5.

59According to OpenAI, the preview received over a million signups within the first five days.

60According to anonymous sources cited by Reuters in December 2022, OpenAI Global, LLC was projecting $200 million of revenue in 2023 and $1 billion in revenue in 2024.

61In January 2023, OpenAI Global, LLC was in talks for funding that would value the company at $29 billion, double its 2021 value.

62On January 23, 2023, Microsoft announced a new $10 billion investment in OpenAI Global, LLC over multiple years, partially needed to use Microsoft's cloud computing service Azure.

63Rumors of this deal suggested that Microsoft may receive 75% of OpenAI's profits until it secures its investment return and a 49% stake in the company.

64The investment is believed to be a part of Microsoft's efforts to integrate OpenAI's ChatGPT into the Bing search engine.

65Google announced a similar AI application, BARD, after ChatGPT was launched, fearing that ChatGPT could threaten Google's place as a go-to source for information.

66On February 7, 2023, Microsoft announced that it was building AI technology based on the same foundation as ChatGPT into Microsoft Bing, Edge, Microsoft 365, and other products.

67On March 3, 2023, Reid Hoffman resigned from his board seat, citing a desire to avoid conflicts of interest with his investments in AI companies via Greylock Partners and his co-founding of the AI startup Inflection AI.

68Hoffman remained on the board of Microsoft, a major investor in OpenAI.

69On March 14, 2023, OpenAI released GPT-4, both as an API with a waitlist and as a feature of ChatGPT+.

70On May 22, 2023, Sam Altman, Greg Brockman, and Ilya Suitskever posted recommendations for the governance of superintelligence.

71They consider that superintelligence could happen within the next 10 years, allowing a dramatically more prosperous future and that given the possibility of existential risk, we can't just be reactive.

72They propose creating an international watchdog organization similar to IAEA to oversee AI systems above a certain capability threshold, suggesting that relatively weak AI systems on the other side should not be overly regulated.

73They also call for more technical safety research for superintelligences and ask for more coordination, for example through governments launching a joint project which many current efforts become part of.

74In July 2023, OpenAI launched the Superalignment Project, aiming to find within four years how to align future superintelligences by automating alignment research using AI.

75In August 2023, it was announced that OpenAI had acquired the New York-based startup Global Illumination, a company that deploys AI to develop digital infrastructure and creative tools.

76On September 21, 2023, Microsoft had begun rebranding all variants of its Copilot to Microsoft Copilot, including the former Bing Chat and the Microsoft 365 Copilot.

77This strategy was followed in December 2023 by adding the MS Copilot to many installations of Windows 11 and Windows 10, as well as a standalone Microsoft Copilot app released for Android and one released for iOS thereafter.

78In October 2023, Sam Altman and Peng Xiao, CEO of the Emirati AI firm G42, announced OpenAI would let G42 deploy OpenAI technology.

79On November 6, 2023, OpenAI launched GPTs, allowing individuals to create customized versions of ChatGPT for specific purposes, further expanding the possibilities of AI applications across various industries.

80On November 14, 2023, OpenAI announced they temporarily suspended new signups for ChatGPT+, due to high demand.

81Access for newer subscribers reopened a month later on December 13, 2024, present.

82Public slash non-profit efforts Sora, partnership with Apple.

83On January 16, 2024, in response to intense scrutiny from regulators around the world, OpenAI announced the formation of a new collective alignment team that would aim to implement ideas from the public for ensuring its models would align to the values of humanity.

84The move was from its public program launched in May 2023.

85The company explained that the program would be separate from its commercial endeavors.

86On January 18, 2024, OpenAI announced a partnership with Arizona State University that would give it complete access to ChatGPT enterprise.

87ASU plans to incorporate the technology into various aspects of its operations, including courses, tutoring, and research.

88It is OpenAI's first partnership with an educational institution.

89In February 2024, the U.S.

90Securities and Exchange Commission was reportedly investigating OpenAI over whether internal company communications made by Altman were used to mislead investors, and an investigation of Altman's statements opened by the southern New York U.S.

91Attorney's office the previous November was ongoing.

92On February 15, 2024, OpenAI announced a text-to-video model named Sora, which it plans to release to the public at an unspecified date.

93It is currently available for red teams for managing critical harms and risks.

94On February 29, 2024, OpenAI and CEO Sam Altman were sued by Elon Musk, who accused them of prioritizing profits over public good, contrary to OpenAI's original mission of developing AI for humanity's benefit.

95The lawsuit cited OpenAI's policy shift after partnering with Microsoft, questioning its open-source commitment and stirring the AI ethics versus profit debate.

96In a blog post, OpenAI stated that Elon understood the mission did not imply open-sourcing AGI.

97In a staff memo, they also denied being a de facto Microsoft subsidiary.

98On March 11, 2024, court filing, OpenAI said it was doing just fine without Elon Musk after he left the company in 2018.

99They also responded to Musk's lawsuit, calling the billionaire's claims incoherent, frivolous, extraordinary, and a fiction.

100On June 11, 2024, Musk unexpectedly withdrew the lawsuit.

101On August 5, 2024, Musk reopened the lawsuit against Altman and others, alleging that Altman claimed that OpenAI was going to be founded as a non-profit organization.

102On May 15, 2024, Ilya Sutskever resigned from OpenAI and was replaced with Jakub Pachaki to be the chief scientist.

103Hours later, Jan Leich, the other co-leader of the super-alignment team, announced his departure, citing an erosion of safety and trust in OpenAI's leadership.

104Their departures, along with several researchers leaving the group, led OpenAI to absorb the team's work into other research areas and officially shut down the super-alignment group.

105According to sources interviewed by Fortune, OpenAI's promise of allocating 20% of its computing capabilities to the super-alignment project had not been fulfilled.

106On May 19, 2024, Reddit and OpenAI announced a partnership to integrate Reddit's content into OpenAI products, including ChatGPT.

107This collaboration allows OpenAI to access Reddit's data API, providing real-time, structured content to enhance AI tools and user engagement with Reddit communities.

108Additionally, Reddit plans to develop new AI-powered features for users and moderators using OpenAI's platform.

109The partnership aligns with Reddit's commitment to privacy, adhering to its public content policy and existing data API terms, which restrict commercial use without approval.

110OpenAI will also serve as a Reddit advertising partner.

111On May 22, 2024, OpenAI entered into an agreement with News Corp to integrate news content from the Wall Street Journal, the New York Post, the Times, and the Sunday Times into its AI platform.

112Meanwhile, other publications like the New York Times chose to sue OpenAI and Microsoft for copyright infringement over the use of their content to train AI models.

113On May 29, 2024, Axios reported that OpenAI had signed deals with Vox Media and The Atlantic to share content to enhance the accuracy of AI models like ChatGPT by incorporating reliable news sources, addressing concerns about AI misinformation.

114Concerns were expressed about the decision by journalists, including those working for the publications, as well as the publications unions.

115The Vox Union stated, as both journalists and workers, we have serious concerns about this partnership, which we believe could adversely impact members of our union, not to mention the well-documented ethical and environmental concerns surrounding the use of generative AI.

116A group of nine current and former OpenAI employees has accused the company of prioritizing profits over safety, using restrictive agreements to silence concerns and moving too quickly with inadequate risk management.

117They call for greater transparency, whistleblower protections, and legislative regulation of AI development.

118On June 10, 2024, it was announced at WWDC 2024 that OpenAI had partnered with Apple Incorporated to bring ChatGPT features to Apple Intelligence and iPhone.

119On June 13, 2024, OpenAI announced that Paul Nakasone, the former head of the NSA, was joining the company's board of directors.

120Nakasone also joined the company's security subcommittee.

121On June 24, 2024, OpenAI acquired Multi, a startup running a collaboration platform based on Zoom.

122In July 2024, Reuters reported that OpenAI was working on a project codenamed Strawberry, previously known as QAsterisk, aiming to enhance AI reasoning capabilities.

123The project reportedly seeks to enable AI to plan ahead, navigate the internet autonomously, and conduct deep research.

124The project was officially released on September 12 and named O1.

125On August 5, TechCrunch reported that OpenAI's co-founder John Shulman has left the company to join rival AI startup Anthropic.

126Shulman cited a desire to focus more deeply on AI alignment research as his reason for the move.

127Additionally, OpenAI's president and co-founder, Greg Brockman, took an extended leave until November.

128In September 2024, OpenAI's global affairs chief, Anna McAnju, expressed support for the UK's approach to AI regulation during her testimony to the House of Lords Communications and Digital Committee, stating that the company favors smart regulation and sees the UK's AI white paper as a positive step towards responsible AI development.

129On September 25, OpenAI's chief technology officer, CTO, Mira Maradi, announced her departure from the company to create the time and space to do my own exploration.

130It had previously been reported Maradi was among those who expressed concerns to the board about Altman.

131In October 2024, OpenAI raised $6.6 billion from investors, potentially valuing the company at $157 billion and solidifying its status as one of the most valuable private firms globally.

132The funding attracted returning venture capital firms like Thrive Capital and Kozla Ventures, along with major backer Microsoft and new investors NVIDIA and SoftBank.

133OpenAI's CFO, Sarah Fryer, informed employees that a tender offer for share buybacks would follow the funding, although specifics were yet to be determined.

134Thrive Capital invested around $1.2 billion, with the option for an additional $1 billion if revenue goals were met.

135Apple, despite initial interest, did not participate in this funding round.

136Also in October 2024, The Intercept revealed that OpenAI's tools were considered essential for AFRICOM's mission and included in an exception to fair opportunity contractual agreement between the Department of Defense and Microsoft.

137In November 2024, OpenAI acquired the long-standing domain chat.com and redirected it to ChatGPT's main site.

138Moreover, Greg Brockman rejoined OpenAI after a three-month leave from his role as president.

139An OpenAI spokesperson confirmed his return, highlighting that Brockman would collaborate with Altman on tackling key technical challenges.

140His return followed a wave of high-profile departures, including Mira Maradi and Ilya Suitskever, who had since launched their own AI ventures.

141Participants, key employees, CEO and co-founder, Sam Altman, former president of the startup accelerator Y Combinator.

142President and co-founder, Greg Brockman, former CTO, third employee of Stripe.

143Chief Scientist Officer, Jakub Pachaki, former director of research at OpenAI.

144Chief Operating Officer, Brad Lightcap, previously at Y Combinator and JPMorgan Chase.

145Chief Financial Officer, Sarah Fryer, former Nextdoor CEO and former CFO at Block Incorporated.

146Chief Product Officer, Kevin Weil, previously at Twitter Incorporated.

147And Meta Platforms, board of directors of the OpenAI non-profit.

148Brett Taylor, chairman.

149Sam Altman, Lawrence Summers, Adam D'Angelo, Sue Desmond Elman, Nicole Seligman, Fiji Simo, Paul Nakasoni, former director of the National Security Agency, 2018-2024.

150Zico Coulter, principal individual investors.

151Reid Hoffman, LinkedIn co-founder.

152Peter Thiel, PayPal co-founder.

153Jessica Livingston, a founding partner of Y Combinator.

154Elon Musk, co-founder.

155Initial motivations.

156Some scientists, such as Stephen Hawking and Stuart Russell, have articulated concerns that if advanced AI gains the ability to redesign itself at an ever-increasing rate, an unstoppable intelligence explosion could lead to human extinction.

157Co-founder Musk characterizes AI as humanity's biggest existential threat.

158Musk and Altman have stated they are partly motivated by concerns about AI safety.

159and the existential risk from artificial general intelligence.

160OpenAI states that it's hard to fathom how much human-level AI could benefit society, and that it is equally difficult to comprehend how much it could damage society if built or used incorrectly.

161Research on safety cannot safely be postponed.

162Because of AI's surprising history, it's hard to predict when human-level AI might come within reach.

163OpenAI states that AI should be an extension of individual human wills and, in the spirit of liberty, as broadly and evenly distributed as possible.

164Co-chair Sam Altman expects the decades-long project to surpass human intelligence.

165Vishal Sikha, former CEO of Infosys, stated that an openness, where the endeavor would produce results generally in the greater interest of humanity, was a fundamental requirement for his support, and that OpenAI aligns very nicely with our long-held values and their endeavor to do purposeful work.

166Cade Metz of Wired suggested that corporations such as Amazon might be motivated by a desire to use open-source software and data to level the playing field against corporations such as Google and Facebook, which own enormous supplies of proprietary data.

167Altman stated that why combinator companies would share their data with OpenAI.

168Strategy.

169In the early years before his 2018 departure, Musk posed the question, What is the best thing we can do to ensure the future is good?

170We could sit on the sidelines or we can encourage regulatory oversight, or we could participate with the right structure with people who care deeply about developing AI in a way that is safe and is beneficial to humanity.

171He acknowledged that there is always some risk that in actually trying to advance friendly AI we may create the thing we are concerned about.

172But nonetheless, that the best defense was to empower as many people as possible to have AI.

173If everyone has AI powers, then there's not any one person or a small set of individuals who can have AI superpower.

174Musk and Altman's counterintuitive strategy, that of trying to reduce the potential harm of AI by giving everyone access to it, is controversial among those concerned with existential risk from AI.

175Philosopher Nick Bostrom said, If you have a button that could do bad things to the world, you don't want to give it to everyone.

176During a 2016 conversation about technological singularity, Altman said, We don't plan to release all of our source code and mentioned a plan to allow wide swaths of the world to elect representatives to a new governance board.

177Greg Brockman stated, Our goal right now is to do the best thing there is to do.

178It's a little vague.

179Conversely, OpenAI's initial decision to withhold GPT-2 around 2019, due to a wish to err on the side of caution in the presence of potential misuse, was criticized by advocates of openness.

180Dilip Rao, an expert in text generation, stated, I don't think spent enough time proving was actually dangerous.

181Other critics argued that open publication was necessary to replicate the research and to create countermeasures.

182More recently, in 2022, OpenAI published its approach to the alignment problem, anticipating that aligning AGI to human values would likely be harder than aligning current AI systems.

183Unaligned AGI could pose substantial risks to humanity and solving the AGI alignment problem could be so difficult that it will require all of humanity to work together.

184They stated that they intended to explore how to better use human feedback to train AI systems and how to safely use AI to incrementally automate alignment research.

185In 2024, following the temporary removal of Sam Altman and his return, many employees gradually left OpenAI, including most of the original leadership team and a significant number of AI safety researchers.

186OpenAI also planned a restructuring to operate as a for-profit company.

187This restructuring could grant Altman a stake in the company.

188Products and Applications Reinforcement Learning At its beginning, OpenAI's research included many projects focused on reinforcement learning, RL.

189OpenAI has been viewed as an important competitor to DeepMind.

190Gym.

191Announced in 2016, Gym is an open-source Python library designed to facilitate the development of reinforcement learning algorithms.

192It aimed to standardize how environments are defined in AI research, making published research more easily reproducible while providing users with a simple interface for interacting with these environments.

193In 2022, new developments of Gym have been moved to the Library Gymnasium.

194Gym Retro.

195Released in 2018, Gym Retro is a platform for reinforcement learning, RL research on video games, using RL algorithms and study generalization.

196Prior RL research focused mainly on optimizing agents to solve single tasks.

197Gym Retro gives the ability to generalize between games with similar concepts but different appearances.

198RoboSumo.

199Released in 2017, RoboSumo is a virtual world where humanoid meta-learning robot agents initially lack knowledge of how to even walk, but are given the goals of learning to move and to push the opposing agent out of the ring.

200Through this adversarial learning process, the agents learn how to adapt to changing conditions.

201When an agent is then removed from this virtual environment and placed in a new virtual environment with high winds, the agent braces to remain upright, suggesting it had learned how to balance in a generalized way.

202Open AI's Igor Mordach argued that competition between agents could create an intelligence arms race that could increase an agent's ability to function even outside the context of the competition.

203Open AI 5.

204Open AI 5 is a team of five Open AI curated bots used in the competitive five-on-five video game Dota 2 that learned to play against human players at a high skill level entirely through trial-and-error algorithms.

205Before becoming a team of five, the first public demonstration occurred at the International 2017, the annual premier championship tournament for the game, where Dendy, a professional Ukrainian player, lost against a bot in a live one-on-one matchup.

206After the match, CTO Greg Brockman explained that the bot had learned by playing against itself for two weeks of real-time, and that the learning software was a step in the direction of creating software that can handle complex tasks like a surgeon.

207The system uses a form of reinforcement learning, as the bots learn over time by playing against themselves hundreds of times a day for months, and are rewarded for actions such as killing an enemy and taking map objectives.

208By June 2018, the ability of the bots expanded to play together as a full team of five, and they were able to defeat teams of amateur and semi-professional players.

209At the International 2018, Open AI 5 played in two exhibition matches against professional players, but ended up losing both games.

210In April 2019, Open AI 5 defeated OG, the reigning world champions of the game at the time, two o'clock in a live exhibition match in San Francisco.

211The bots' final public appearance came later that month, where they played in 42,729 total games in a four-day open online competition, winning 99.4% of those games.

212Open AI 5's mechanisms in Dota 2's bot player shows the challenges of AI systems in multiplayer online battle arena, MOBA games, and how Open AI 5 has demonstrated the use of deep reinforcement learning, DRL, agents to achieve superhuman competence in Dota 2 matches.

213Dactyl.

214Developed in 2018, Dactyl uses machine learning to train a shadow hand, a human-like robot hand, to manipulate physical objects.

215It learns entirely in simulation using the same RL algorithms and training code as Open AI 5.

216Open AI tackled the object orientation problem by using domain randomization, a simulation approach which exposes the learner to a variety of experiences rather than trying to fit to reality.

217The setup for Dactyl, aside from having motion tracking cameras, also has RGB cameras to allow the robot to manipulate an arbitrary object by seeing it.

218In 2018, Open AI showed that the system was able to manipulate a cube and an octagonal prism.

219In 2019, Open AI demonstrated that Dactyl could solve a Rubik's Cube.

220The robot was able to solve the puzzle 60% of the time.

221Objects like the Rubik's Cube introduce complex physics that is harder to model.

222Open AI did this by improving the robustness of Dactyl to perturbations by using automatic domain randomization, ADR.

223A simulation approach of generating progressively more difficult environments.

224ADR differs from manual domain randomization by not needing a human to specify randomization ranges.

225Op B.

226In June 2020, Open AI announced a multipurpose API which it said was for accessing new AI models developed by Open AI to let developers call on it for any English-language AI task.

227Text generation.

228The company has popularized generative pre-trained transformers.

229GPT.

230Open AI's original GPT model.

231GPT-1.

232The original paper on generative pre-training of a transformer-based language model was written by Alec Radford and his colleagues, and published in pre-print on Open AI's website on June 11, 2018.

233It showed how a generative model of language could acquire world knowledge and process long-range dependencies by pre-training on a diverse corpus with long stretches of contiguous text.

234GPT-2.

235Generative pre-trained transformer-2.

236GPT-2.

237Is an unsupervised transformer language model and the successor to Open AI's original GPT model, GPT-1.

238GPT-2 was announced in February 2019, with only limited demonstrative versions initially released to the public.

239The full version of GPT-2 was not immediately released due to concern about potential misuse, including applications for writing fake news.

240Some experts expressed skepticism that GPT-2 posed a significant threat.

241In response to GPT-2, the Allen Institute for Artificial Intelligence responded with a tool to detect neural fake news.

242Other researchers, such as Jeremy Howard, warned of the technology to totally fill Twitter, email, and the web up with reasonable-sounding, context-appropriate prose, which would drown out all other speech and be impossible to filter.

243In November 2019, Open AI released the complete version of the GPT-2 language model.

244Several websites host interactive demonstrations of different instances of GPT-2 and other transformer models.

245GPT-2's authors argue unsupervised language models to be general-purpose learners, illustrated by GPT-2 achieving state-of-the-art accuracy and perplexity on seven-of-eight zero-shot tasks.

246That is, the model was not further trained on any task-specific input-output examples.

247The corpus it was trained on, called WebText, contains slightly 40 gigabytes of text from URLs shared in Reddit submissions with at least three upvotes.

248It avoids certain issues encoding vocabulary with word tokens by using byte-pair encoding.

249This permits representing any string of characters by encoding both individual characters and multiple-character tokens.

250GPT-3, first described in May 2020, generative pre-trained Transformer 3, GPT-3, is an unsupervised transformer language model and the successor to GPT-2.

251Open AI stated that the full version of GPT-3 contained 175 billion parameters, two orders of magnitude larger than the 1.5 billion in the full version of GPT-2, although GPT-3 models with as few as 125 million parameters were also trained.

252Open AI stated that GPT-3 succeeded at certain meta-learning tasks and could generalize the purpose of a single input-output pair.

253The GPT-3 release paper gave examples of translation and cross-linguistic transfer learning between English and Romanian and between English and German.

254GPT-3 dramatically improved benchmark results over GPT-2.

255Open AI cautioned that such scaling up of language models could be approaching or encountering the fundamental capability limitations of predictive language models.

256Pre-training GPT-3 required several thousand petaflops slash s-days of compute, compared to tens of petaflops slash s-days for the full GPT-2 model.

257Like its predecessor, the GPT-3 trained model was not immediately released to the public for concerns of possible abuse, although Open AI planned to allow access through a paid cloud API after a two-month free private beta that began in June 2020.

258On September 23, 2020, GPT-3 was licensed exclusively to Microsoft.

259Codex.

260Announced in mid-2021, Codex is a descendant of GPT-3 that has additionally been trained on code from 54 million GitHub repositories and is the AI powering the code autocompletion tool GitHub co-pilot.

261In August 2021, an API was released in private beta.

262According to Open AI, the model can create working code in over a dozen programming languages most effectively in Python.

263Several issues with glitches, design flaws, and security vulnerabilities were cited.

264GitHub co-pilot has been accused of emitting copyrighted code with no author attribution or license.

265Open AI announced that they would discontinue support for Codex API on March 23, 2023.

266GPT-4.

267On March 14, 2023, Open AI announced the release of generative pre-trained Transformer 4, GPT-4, capable of accepting text or image inputs.

268They announced that the updated technology passed a simulated law school bar exam with a score around the top 10% of test takers.

269By contrast, GPT-3.5 scored around the bottom 10%.

270They said that GPT-4 could also read, analyze, or generate up to 25,000 words of text and write code in all major programming languages.

271Observers reported that the iteration of ChatGPT using GPT-4 was an improvement on the previous GPT-3.5.

272Based iteration, with the caveat that GPT-4 retained some of the problems with earlier revisions.

273GPT-4 is also capable of taking images as input on ChatGPT.

274Open AI has declined to reveal various technical details and statistics about GPT-4, such as the precise size of the model.

275GPT-4.

276On May 13, 2024, Open AI announced and released GPT-4-0, which can process and generate text, images, and audio.

277GPT-4-0 achieved state-of-the-art results in voice, multilingual, and vision benchmarks, setting new records in audio speech recognition and translation.

278It scored 88.7% on the Massive Multitask Language Understanding, MMLU, benchmark compared to 86.5% by GPT-4.

279On July 18, 2024, Open AI released GPT-4-0 Mini, a smaller version of GPT-4-0 replacing GPT-3.5 Turbo on the ChatGPT interface.

280Its API costs $0.15 per million input tokens and $0.60 per million output tokens, compared to $5 and $15 respectively for GPT-4-0.

281Open AI expects it to be particularly useful for enterprises, startups, and developers seeking to automate services with AI agents.

282O-1.

283On September 12, 2024, Open AI released the O-1 Preview and O-1 Mini models, which have been designed to take more time to think about their responses, leading to higher accuracy.

284These models are particularly effective in science, coding, and reasoning tasks, and were made available to ChatGPT Plus and team members.

285Image Classification CLIP, revealed in 2021, CLIP, Contrastive Language Image Pre-Training, is a model that is trained to analyze the semantic similarity between text and images.

286It can notably be used for image classification.

287Text to image.

288Dale.

289Revealed in 2021, DALL-E is a transformer model that creates images from textual descriptions.

290DALL-E uses a 12 billion parameter version of GPT-3 to interpret natural language inputs, such as a green leather purse shaped like a pentagon or an isometric view of a sad capybara, and generate corresponding images.

291It can create images of realistic objects, a stained glass window with an image of a blue strawberry, as well as objects that do not exist in reality, a cube with the texture of a porcupine.

292As of March 2021, no API or code is available.

293Dale.

294In April 2022, OpenAI announced DALL-E2, an updated version of the model with more realistic results.

295In December 2022, OpenAI published on GitHub software for PointE, a new rudimentary system for converting a text description into a three-dimensional model.

296Dale-3.

297In September 2023, OpenAI announced DALL-E3, a more powerful model better able to generate images from complex descriptions without manual prompt engineering and render complex details like hands and text.

298It was released to the public as a ChatGPT Plus feature in October.

299Text to video.

300Sora.

301Sora is a text video model that can generate videos based on short descriptive prompts, as well as extend existing videos forwards or backwards in time.

302It can generate videos with resolution up to 1920 by 1080 or 1080 by 1920.

303The maximal length of generated videos is unknown.

304Sora's development team named it after the Japanese word for sky to signify its limitless creative potential.

305Sora's technology is an adaptation of the technology behind the DALL-E3 text-to-image model.

306OpenAI trained the system using publicly available videos as well as copyrighted videos licensed for that purpose, but did not reveal the number or the exact sources of the videos.

307OpenAI demonstrated some Sora-created high-definition videos to the public on February 15, 2024, stating that it could generate videos up to one minute long.

308It also shared a technical report highlighting the methods used to train the model and the model's capabilities.

309It acknowledged some of its shortcomings, including struggles simulating complex physics.

310Well, Douglas Heaven of the MIT Technology Review called the demonstration videos impressive, but noted that they must have been cherry-picked and might not represent Sora's typical output.

311Despite skepticism from some academic leaders following Sora's public demo, notable entertainment industry figures have shown significant interest in the technology's potential.

312In an interview, actor-slash-filmmaker Tyler Perry expressed his astonishment at the technology's ability to generate realistic video from text descriptions, citing its potential to revolutionize storytelling and content creation.

313He said that his excitement about Sora's possibilities was so strong that he had decided to pause plans for expanding his Atlanta-based movie studio.

314Speech-to-text Whisper.

315Released in 2022, Whisper is a general-purpose speech recognition model.

316It is trained on a large dataset of diverse audio and is also a multitask model that can perform multilingual speech recognition as well as speech translation and language identification.

317Music generation.

318Musenet.

319Released in 2019, Musenet is a deep neural net trained to predict subsequent musical notes in MIDI music files.

320Can generate songs with 10 instruments in 15 styles.

321According to The Verge, a song generated by Musenet tends to start reasonably, but then fall into chaos the longer it plays.

322In pop culture, initial applications of this tool were used as early as 2020 for the internet psychological thriller Ben Drown to create music for the titular character.

323Jukebox.

324Released in 2020, Jukebox is an open-sourced algorithm to generate music with vocals.

325After training on 1.2 million samples, the system accepts a genre, artist, and a snippet of lyrics and output song samples.

326OpenAI stated the songs show local musical coherence follow traditional chord patterns, but acknowledged that the songs lack familiar larger musical structures, such as choruses that repeat, and that there is a significant gap between jukebox and human-generated music.

327The Verge stated it's technologically impressive, even if the results sound like mushy versions of songs that might feel familiar.

328While Business Insider stated surprisingly, some of the resulting songs are catchy and sound legitimate.

329User Interfaces.

330Debate Game.

331In 2018, OpenAI launched the Debate Game, which teaches machines to debate toy problems in front of a human judge.

332The purpose is to research whether such an approach may assist in auditing AI decisions and in developing explainable AI.

333Microscope.

334Released in 2020, Microscope is a collection of visualizations of every significant layer and neuron of eight neural network models which are often studied in interpretability.

335Microscope was created to analyze the features that form inside these neural networks easily.

336The models included are AlexNet, VGG19, different versions of Inception, and different versions of ClipResNet.

337ChatGPT.

338Launched in November 2022, ChatGPT is an artificial intelligence tool built on top of GPT-3 that provides a conversational interface that allows users to ask questions in natural language.

339The system then responds with an answer within seconds.

340ChatGPT reached one million users five days after its launch.

341As of 2023, ChatGPT Plus is a GPT-4-backed version of ChatGPT available for a US $20 per month subscription fee.

342The original version is backed by GPT-3.5.

343OpenAI also makes GPT-4 available to a select group of applicants through their GPT-4 API waitlist.

344After being accepted, an additional fee of $0.03 per 1,000 tokens in the initial text provided to the model prompt and $0.06 per 1,000 tokens that the model generates completion is charged for access to the version of the model with an 8,192 token context window.

345For the 32,768 token context window, the prices are doubled.

346In May 2023, OpenAI launched a user interface for ChatGPT for the App Store on iOS and later in July 2023 for the Play Store on Android.

347The app supports chat history syncing and voice input using Whisper, OpenAI's speech recognition model.

348In September 2023, OpenAI announced that ChatGPT can now see, hear, and speak.

349ChatGPT Plus users can upload images, while mobile app users can talk to the chatbot.

350In October 2023, OpenAI's latest image generation model, DALL-E3, was integrated into ChatGPT Plus and ChatGPT Enterprise.

351The integration uses ChatGPT to write prompts for DALL-E guided by conversation with users.

352OpenAI's GPT Store, initially slated for a 2023 launch, is now deferred to an undisclosed date in early 2024, attributed likely to the leadership changes in November following the initial announcement.

353Concerns about the energy consumption of generative AI, including ChatGPT, are rising.

354In September 2024, Microsoft entered a deal with Constellation Energy to reopen the Three Mile Island nuclear plant to supply power to its AI-driven data centers.

355SearchGPT SearchGPT, a prototype search engine developed by OpenAI, was unveiled on July 25, 2024, with an initial limited release to 10,000 test users.

356It combines traditional search engine features with generative AI capabilities.

357Stargate and other supercomputers.

358Stargate is a potential artificial intelligence supercomputer in development by Microsoft and OpenAI.

359Stargate is designed as part of a greater data center project, which could represent an investment of as much as $100 billion by Microsoft.

360Stargate is reported to be part of a series of AI-related construction projects planned in the next few years by the companies Microsoft and OpenAI.

361The supercomputers will be constructed in five phases.

362The fourth phase should consist in a smaller OpenAI supercomputer, planned to launch around 2026.

363Stargate is the fifth and final phase of the program, and will take five and six years to complete and is slated to launch around 2028.

364The artificial intelligence of Stargate is slated to be contained on millions of special server chips.

365The supercomputers data center will be built in the U.S.

366across 700 acres of land.

367It has a planned power consumption of five gigawatts, for which it could rely on nuclear energy.

368The name Stargate is a homage to the 1994 sci-fi film Stargate.

369Controversies.

370Firing of Altman.

371On November 17, 2023, Sam Altman was removed as CEO when its board of directors, composed of Helen Toner, Ilya Sutskever, Adam D'Angelo, and Tasha McCauley, cited a lack of confidence in him.

372Chief Technology Officer Mira Maradi took over as interim CEO.

373Greg Brockman, the president of OpenAI, was also removed as chairman of the board and resigned from the company's presidency shortly thereafter.

374Three senior OpenAI researchers subsequently resigned.

375Director of Research and GPT-4 lead Jakub Pachaki, head of AI Risk Alexander Madri, and researcher Simon Sitter.

376On November 18, 2023, there were reportedly talks of Altman returning as CEO amid pressure placed upon the board by investors such as Microsoft and Thrive Capital, who objected to Altman's departure.

377Although Altman himself spoke in favor of returning to OpenAI, he has since stated that he considered starting a new company and bringing former OpenAI employees with him if talks to reinstate him didn't work out.

378The board members agreed in principle to resign if Altman returned.

379On November 19, 2023, negotiations with Altman to return failed and Maradi was replaced by Emmett Shear as interim CEO.

380The board initially contacted Anthropic CEO Dario Amadei, a former OpenAI executive, about replacing Altman, and proposed a merger of the two companies, but both offers were declined.

381On November 20, 2023, Microsoft CEO Satya Nadella announced Altman and Brockman would be joining Microsoft to lead a new advanced AI research team, but added that they were still committed to OpenAI despite recent events.

382Before the partnership with Microsoft was finalized, Altman gave the board another opportunity to negotiate with him.

383About 738 of OpenAI's 770 employees, including Maradi and Suitskever, signed an open letter stating they would quit their jobs and join Microsoft if the board did not rehire Altman and then resign.

384This prompted OpenAI investors to consider legal action against the board as well.

385In response, OpenAI management sent an internal memo to employees stating that negotiations with Altman and the board had resumed and would take some time.

386On November 21, 2023, after continued negotiations, Altman and Brockman returned to the company in their prior roles along with a reconstructed board made up of new members Brett Taylor as chairman and Lawrence Summers, with D'Angelo remaining.

387On November 22, 2023, emerging reports suggested that Sam Altman's dismissal from OpenAI may have been linked to his alleged mishandling of a significant breakthrough in the organization's secretive project codenamed QSTAR.

388According to sources within OpenAI, QAsterisk is aimed at developing AI capabilities in logical and mathematical reasoning and reportedly involves performing math on the level of grade school students.

389Concerns about Altman's response to this development, specifically regarding the discovery's potential safety implications, were reportedly raised with the company's board shortly before Altman's firing.

390On November 29, 2023, OpenAI announced that an anonymous Microsoft employee had joined the board as a non-voting member to observe the company's operations.

391Microsoft resigned from the board in July 2024.

392Content moderation contract with Sama.

393In January 2023, OpenAI has been criticized for outsourcing the annotation of datasets to Sama, a company based in San Francisco that employed workers in Kenya.

394These annotations were used to train an AI model to detect toxicity, which could then be used to moderate toxic content, notably from ChatGPT's training data and outputs.

395However, these pieces of text usually contained detailed descriptions of various types of violence, including sexual violence.

396The investigation uncovered that OpenAI began sending snippets of data to Sama as early as November 2021.

397The four Sama employees interviewed by Time described themselves as mentally scarred.

398OpenAI paid Sama $12.50 per hour of work, and Sama was redistributing the equivalent of between $1.32 and $2 per hour post-tax to its annotators.

399Sama's spokesperson said that the $12.50 was also covering other implicit costs, among which were infrastructure expenses, quality assurance, and management.

400Lack of technological transparency.

401In March 2023, the company was also criticized for disclosing particularly few technical details about products like GPT-4, contradicting its initial commitment to openness and making it harder for independent researchers to replicate its work and develop safeguards.

402OpenAI cited competitiveness and safety concerns to justify this strategic turn.

403OpenAI's former chief scientist Ilya Suitskever argued in 2023 that open-sourcing increasingly capable models was increasingly risky, and that the safety reasons for not open-sourcing the most potent AI models would become obvious in a few years.

404Non-disparagement agreement.

405On May 17, 2024, a Vox article reported that OpenAI was asking departing employees to sign a lifelong non-disparagement agreement, forbidding them from criticizing OpenAI or acknowledging the existence of the agreement.

406Daniel Cocotaglio, a former employee, publicly stated that he forfeited his vested equity in OpenAI in order to leave without signing the agreement.

407Sam Altman stated that he was unaware of the equity cancellation provision, and that OpenAI never enforced it to cancel any employees' vested equity.

408Vox published leaked documents and emails challenging this claim.

409On May 23, 2024, OpenAI sent a memo releasing former employees from the agreement.

410Copyright infringement in training data.

411OpenAI was sued for copyright infringement by authors Sarah Silverman, Matthew Butterick, Paul Tremblay, and Mona Awad in July 2023.

412In September 2023, 17 authors, including George R.

413R.

414Martin, John Grisham, Jody Picoult, and Jonathan Franson, joined the Authors Guild in filing a class action lawsuit against OpenAI, alleging that the company's technology was illegally using their copyrighted work.

415The New York Times also sued the company in late December 2023.

416In May 2024, it was revealed that OpenAI had destroyed its Books 1 and Books 2 training datasets, which were used in the training of GPT-3, and which the Authors Guild believed to have contained over 100,000 copyrighted books.

417In 2021, OpenAI developed a speech recognition tool called Whisper.

418OpenAI used it to transcribe more than 1 million hours of YouTube videos into text for training GPT-4.

419The automated transcription of YouTube videos raised concerns within OpenAI employees regarding potential violations of YouTube's terms of service, which prohibit the use of videos for applications independent of the platform, as well as any type of automated access to its videos.

420Despite these concerns, the project proceeded with notable involvement from OpenAI's president, Greg Brockman.

421The resulting dataset proved instrumental in training GPT-4.

422In February 2024, The Intercept as well as Raw Story and Alternate Media Incorporated filed lawsuit against OpenAI on copyright litigation ground.

423The lawsuit is said to have charted a new legal strategy for digital-only publishers to sue OpenAI.

424On April 30, 2024, eight newspapers filed a lawsuit in the Southern District of New York against OpenAI and Microsoft claiming illegal harvesting of their copyrighted articles.

425The suing publications included The Mercury News, The Denver Post, The Orange County Register, ST, Paul Pioneer Press, Chicago Tribune, Orlando Sentinel, Sun Sentinel, and New York Daily News.

426GDPR Compliance.

427In April 2023, the EU's European Data Protection Board, EDPB, formed a dedicated task force on ChatGPT to foster cooperation and to exchange information on possible enforcement actions conducted by data protection authorities based on the enforcement action undertaken by the Italian Data Protection Authority against OpenAI about the ChatGPT service.

428In late April 2024, NOYB filed a complaint with the Austrian Datenschutzbehord against OpenAI for violating the European General Data Protection Regulation.

429A text created with ChatGPT gave a false date of birth for a living person without giving the individual the option to see the personal data used in the process.

430A request to correct the mistake was denied.

431Additionally, neither the recipients of ChatGPT's work nor the sources used could be made available, OpenAI claimed.

432Removal of Military and Warfare Clause.

433OpenAI quietly deleted its ban on using ChatGPT for military and warfare.

434Up until January 10, 2024, its usage policies included a ban on activity that has high risk of physical harm, including specifically weapons development and military and warfare.

435Its new policies prohibit our service to harm yourself or others and to develop or use weapons.

436As one of the industry collaborators, OpenAI provides LLM to the Artificial Intelligence Cyber Challenge, A9CC, sponsored by Defense Advanced Research Projects Agency, DARPA, and Advanced Research Projects Agency for Health to Protect Software Critical to Americans.

437In October 2024, The Intercept revealed that OpenAI's tools are considered essential for AFRICOM's mission and included in an exception to Fair Opportunity Contract Ural Agreement between the United States Department of Defense and Microsoft.

438Use in state-backed influence operations.

439In May 2024, OpenAI removed accounts used by state-backed influence operations such as China's Spamoflage and Russia's Doppelganger.

440OpenAI banned access to its services from China in July 2024.

441Data Scraping.

442In June 2023, a lawsuit claimed that OpenAI scraped 300 billion words online without consent and without registering as a data broker.

443It was filed in San Francisco, California, by 16 anonymous plaintiffs.

444They also claimed that OpenAI and its partner, as well as customer Microsoft, continued to unlawfully collect and use personal data from millions of consumers worldwide to train artificial intelligence models.

445This was the English Wikipedia article about OpenAI as of December 4th, 2024 narrated by AI-generated voice via the open-source software Sony Translate.