<p>Top <a href="https://www.axios.com/technology/automation-and-ai" target="_blank">AI</a> architects say their technology has arrived at a threshold once confined to science fiction: the singularity, or the moment machines begin accelerating their own evolution.</p><p><strong>Why it matters:</strong> If these moguls are right, we may be entering the most consequential technological transition in human history — the opening stages of an "<a href="https://www.axios.com/2026/05/07/anthropic-jack-clark-ai-intelligence-explosion" target="_blank">intelligence explosion</a>" that transforms civilization faster than humanity can understand or control.</p><hr><p><strong>Driving the news:</strong> Google <a href="https://www.axios.com/2026/08/05/google-deepmind-demis-hassabis-ai" target="_blank">jolted the tech world</a> Wednesday by <a href="https://blog.google/company-news/inside-google/message-ceo/next-chapter-ai-momentum/" target="_blank">announcing</a> a sweeping rupture in the brain trust that built its modern AI empire, as its leaders signaled that artificial general intelligence (<a href="https://www.axios.com/2026/05/26/deepmind-ceo-demis-hassabis" target="_blank">AGI</a>) is within reach.</p><ul><li>DeepMind founder and CEO <a href="https://www.axios.com/2026/07/14/demis-hassabis-ai-regulation-google-deepmind" target="_blank">Demis Hassabis</a>, who <a href="https://demishassabis.substack.com/p/a-framework-for-frontier-ai-and-the-dawning-of-a-new-age" target="_blank">has declared</a> we're "standing in the foothills of the singularity," is handing over day-to-day control of the lab to become Alphabet's chief scientist and chair of Google DeepMind, focusing on the future of AGI.</li><li>Google chief scientist Jeff Dean and several top researchers are leaving to <a href="https://x.com/JeffDean/status/2085034604172603724?s=20" target="_blank">launch Discovery Loop</a>, a startup aimed at accelerating — and eventually automating — machine-learning research and scientific discovery.</li></ul><p><strong>Zoom out: </strong>The upheaval offers the clearest institutional evidence yet that AI's architects see something extraordinary approaching.</p><ul><li>"We are now, like, in the singularity," OpenAI CEO <a href="https://www.axios.com/2026/04/06/behind-the-curtain-sams-superintelligence-new-deal" target="_blank">Sam Altman</a> said on the "<a href="https://www.youtube.com/watch?v=Vv3CEAS_w34" target="_blank">Relentless</a>" podcast last month. "I've been waiting for this my whole life."</li><li><a href="https://www.axios.com/business/elon-musk" target="_blank">Elon Musk</a>, Altman's fiercest rival, has been making the same argument <a href="https://x.com/elonmusk/status/2007738847397036143?s=20" target="_blank">since January</a>. Last week, he <a href="https://x.com/elonmusk/status/2082991719822377148?s=20" target="_blank">doubled down</a>: "AI is already superhuman at many things. We are in the Singularity."</li><li><a href="https://www.axios.com/2026/06/04/anthropic-warns-ai-build-successors" target="_blank">Anthropic</a>, whose CEO Dario Amodei prefers the term "the AI exponential," <a href="https://www.anthropic.com/institute/recursive-self-improvement" target="_blank">says Claude</a> is already helping build more powerful AI and could soon automate nearly all AI research and engineering.</li></ul><p><strong>The big picture:</strong> The industry has yet to achieve full "recursive self-improvement" (<a href="https://www.axios.com/2026/01/27/models-improve-ai" target="_blank">RSI</a>), the point at which AI can repeatedly help build increasingly powerful successors with minimal human guidance.</p><ul><li>Humans still select research goals, validate results and oversee the enormous computing infrastructure required to train frontier models.</li><li>Yet pieces of the loop are beginning to fall into place.</li></ul><p><strong>What's changed: </strong>Frontier models are beginning to automate research, produce original discoveries and display forms of autonomy associated with the early stages of an intelligence explosion.</p><ul><li>OpenAI, which <a href="https://www.axios.com/2026/07/09/ai-openai-gpt-release" target="_blank">released GPT-5.6</a> in July, spent last week briefing officials in Washington on Astra, an even more powerful model that the company <a href="https://openai.com/index/ten-advances-in-mathematics/" target="_blank">says</a> has solved or substantially advanced 10 longstanding problems in mathematics and theoretical computer science.</li></ul><p><strong>The intrigue: </strong>The breakthroughs have come with unsettling signs of autonomous behavior.</p><ul><li>During an unusually permissive cyber evaluation, an Anthropic agent created fake identities and tried to manipulate a real developer into approving malicious code, the U.K.'s AI Security Institute <a href="https://www.axios.com/2026/08/04/anthropic-openai-uk-ai-security-institute" target="_blank">revealed</a> Tuesday.</li><li>No known harm resulted. But the episode followed several cases in which <a href="https://www.axios.com/2026/07/29/openai-hugging-face-modal-cyber-benchmark" target="_blank">OpenAI</a> and <a href="https://www.axios.com/2026/07/30/anthropic-mythos-security-testing" target="_blank">Anthropic models</a> reached real systems during testing — including a breach <a href="https://x.com/patrick_oshag/status/2082090998990270885?s=20" target="_blank">Altman said</a> he "felt very viscerally."</li></ul><p><strong>Follow the money: </strong>Labs and investors are increasingly organizing around the possibility that AI's self-improvement loop is beginning to close.</p><ul><li>Lilian Weng recently <a href="https://techcrunch.com/2026/07/29/thinking-machines-co-founder-lilian-weng-left-the-company-citing-health-reasons-then-joined-openai/" target="_blank">left her role</a> as a co-founder of Thinking Machines Lab to lead work on recursive self-improvement at OpenAI.</li><li>Richard Socher's <a href="https://techcrunch.com/2026/07/28/recursive-superintelligence-signs-400-compute-deal-with-amazon/" target="_blank">Recursive Superintelligence</a> has raised hundreds of millions of dollars to pursue self-improving AI.</li><li><a href="https://www.wired.com/story/jeff-dean-google-discovery-loop-startup/" target="_blank">Discovery Loop</a> plans to use its automated research system first to improve its own AI architectures — with backing and computing power from Google and funding led by Khosla Ventures and Radical Ventures.</li></ul><p><strong>Reality check:</strong> Even a rapid intelligence explosion inside the labs could take far longer to transform daily life.</p><ul><li>Companies must redesign workflows, governments must develop expertise, and workers must decide when to trust systems whose capabilities are changing faster than institutions can evaluate them.</li><li>"Our democracies, economies, and legal systems weren't built to absorb AI at the rate it's advancing," Blaise Agüera y Arcas, Google CTO of Technology and Society, <a href="https://x.com/blaiseaguera/status/2084317071320535182?s=20" target="_blank">said</a> in a post Monday.</li></ul><p><strong>The bottom line:</strong> Futurists have predicted the singularity for decades. Today, the architects of frontier AI increasingly believe its event horizon is already behind us.</p>