Additionally, OpenAI is working on a “super app.”
This is the latest issue of The Download, our weekday newsletter offering daily updates on technological developments.
OpenAI is committed to developing a completely automated research assistant
OpenAI has set a new ambitious target: creating an AI researcher—a fully autonomous agent-based system capable of independently addressing complex, large-scale challenges. The San Francisco company has declared this objective as its “north star” for the coming years.
By September, the organization plans to develop “an independent AI research intern” that will handle a limited number of defined research issues. This intern will act as a precursor to the comprehensive automated multi-agent system expected to launch in 2028.
In an exclusive conversation this week, OpenAI’s chief scientist, Jakub Pachocki, outlined the plans. Discover what I found out.
—Will Douglas Heaven
Psychedelic substances are (still) underperforming in clinical studies
In the past ten years, we have witnessed a surge of scientific interest in psychedelic drugs. Compounds such as psilocybin—found in magic mushrooms—are being investigated for various health applications, including treating depression, PTSD, addiction, and even obesity. However, two studies published earlier this week illustrate the challenges involved in researching these substances.
To me, they highlight the extent to which these drugs have been overpromised. Learn why here.
—Jessica Hamzelou
This article originally appeared in The Checkup, MIT Technology Review’s weekly biotechnology newsletter. Subscribe to have it delivered to your inbox every Wednesday.
Read more: What effects do psychedelic drugs have on our brains? AI could assist in discovering
The essential reads
I’ve scoured the internet to bring you today’s most engaging/important/scary/captivating tales from the world of technology.
1 OpenAI is developing a “super app”
It’s consolidating ChatGPT, a web browser, and a programming tool into one application. (The Verge)
+ It’s also acquiring coding startup Astral to improve its Codex model. (Ars Technica)
+ These initiatives come as the company scales back on side ventures. (WSJ $)
+ OpenAI has fallen behind Anthropic in the enterprise sector. (Axios)
2 The US has charged Super Micro’s co-founder with smuggling AI technologies to China
Super Micro ranks third on Fortune’s fastest-growing companies list. (Reuters)
+ GenAI is acquiring skills for espionage on behalf of the US military. (MIT Technology Review)
+ The competition for computing resources is influencing the China-US dynamics. (Politico)
3 The DoJ has dismantled the botnets responsible for the largest DDoS attack in history
They had compromised over 3 million devices. (Wired $)
+ The DoJ has also taken control of domains belonging to Iranian “hacktivists.” (Axios)
4 The Pentagon claims that Anthropic’s foreign workforce poses security threats
It specifically mentioned concerns about employees from China. (Axios)
+ Anthropic’s ethical limits have frustrated the DoD. (MIT Technology Review)
5 Rising oil prices could jeopardize the AI industry, warns the WTO
Concerns are increasing regarding a potential energy crisis. (The Guardian)
+ We analyzed AI’s energy consumption. (MIT Technology Review)
6 Jeff Bezos is attempting to raise $100 billion to implement AI in manufacturing
The capital would be used to purchase manufacturing companies and integrate them with AI. (WSJ $)
+ Read about how to tailor AI for success. (MIT Technology Review)
7 The creator of Signal is assisting in securing Meta’s AI
Moxie Marlinspike is integrating his encrypted chatbot, Confer. (Wired $)
+ Meta is again relying on AI instead of human moderators. (CNBC)
+ AI is facilitating online criminal activity. (MIT Technology Review)
8 Prediction market Kalshi has secured $1 billion at a valuation of $22 billion
That’s twice its valuation from December. (Bloomberg $)
+ Arizona’s Attorney General has filed charges against the company for “illegal gambling.” (NPR)
9 Meta isn’t shutting down Horizon Worlds for VR after all
It has reversed plans to terminate the metaverse app (for now). (CNBC)
10 A US startup is seeking an “AI bully”
The chosen applicant must test the limits of leading chatbots. (The Guardian)
Quote of the day
“Picture a sports bar… just for situational monitoring — live X feeds, flight tracking, Bloomberg terminals, and Polymarket screens.”
—Kalshi competitor Polymarket reveals its dystopian concept for a new establishment.
One More Thing
The rise of gamification across the globe
It’s a thought that strikes every gamer at some point: what if the odd, intensely focused state experienced in video games could somehow be applied to the real world?
For a select group of consultants, startup visionaries, and game designers during the late 2000s, this state of “euphoric productivity” was recognized as the key to realizing our inherent human capabilities. This vision gave rise to the widespread phenomenon of gamification—but it failed to deliver on its promises.
Rather than setting us free, gamification evolved into a mechanism for manipulation, distractions, and oversight. Find out why we embraced it—and how we might reclaim our autonomy.
—Bryan Gardiner
We can still enjoy nice things
A space for comfort, enjoyment, and distraction to enhance your day. (Share your ideas? Contact me.)
+ In a significant legal victory for trolling, Afroman has prevailed in his diss track lawsuit against the police.
+ This LEGO artist transforms standard sets into completely unique iconic creations.
+ Discover aliens with these interactive estimates of advanced civilizations.
+ A rare superbloom in Death Valley has been captured on camera.
