This issue of WDODTW? is brought to you free thanks to POTATO.
If you’re in UK tech, you’ll want to be a member of POTATO, the underground network that’s saving members bank on top events including SXSW, London Tech Week, AI Summit and more. +13 channels only on WhatsApp bring members critical thinking, analysis, events, and more. Spaces are limited (thank Zuck).
Join over at potato.tech.
OpenAI Built For Science. Codex Ate The Rest.
OpenAI focused instead of launching everything at once. GPT-Rosalind is aimed squarely at drug research, not as a general tool trying to fit every industry. Drug discovery needs reliable results, access to specialist data and deep links into existing lab systems. Most broad AI models struggle with that level of precision.
Working with companies like Moderna and Amgen shows this is meant to be used inside real research pipelines, not just shown in demos. Choosing a narrow, high value field makes it harder for competitors to copy quickly. Industries with strict rules, long timelines and expensive mistakes reward systems that are deeply integrated and dependable, not ones that simply add new features every few months.
A drastically updated Codex moved in the opposite direction, going wider, deeper and more invasive. Computer control, browser use, file inspection and cross tool execution reposition it from coding assistant to execution layer. A clear swipe at Anthropic’s recent media wins. Slack, Notion and internal systems are now pulled into one interface. Memory, plugins and Agents SDK extend reach into long running tasks inside controlled environments so you go from chat to action faster. Work no longer passes through prompts alone. Work runs inside a system that coordinates context, state and permissions. Switching cost begins accumulating at the workflow layer, not the model layer. In short, this is a win for enterprise.
Leadership exits are reinforcing (fuelling?) chatter of more consolidation too. Fewer parallel bets, and a tighter commercial tone per Denise Dresser’s memo which name checked the competition. Trusted Access expanded into security and defence contexts while SoftBank reinforced capital structure. Data centre delays reveal physical constraints behind digital ambition. Public scrutiny around governance and Sam Altman’s position adds narrative volatility at the centre of capital gravity.
SO WHAT?
There’s lots going on with OpenAI right now, multiple fires (literally and PR-based) tp put out. Product moves continue to add vertical depth, but OpenAI looks like they are losing ground to Anthropic at every turn. The key insight this week came from a rare PR slip, you never mention the competition and this week senior execs did just that.
One move goes deep into specialised areas like drug research, where accuracy and tight partnerships matter and where copying is hard. Another move goes wide by turning Codex into a system that actually carries out tasks across your apps, not just suggests text or code.
Pressure is building while all this is happening which is likely impacting internal elements (staff, shipping, scrutiny, paranoia). Moving into defence and regulated industries attracts external political and regulatory scrutiny. OpenAI is no longer testing ideas in public and a lot feels reactive (whether or not that’s the case - every AI company out there is moving at speeds no-one is comfortable with). For OpenAI product focus, capital backing and infrastructure are now aligned around owning both specialised high value sectors and the space where everyday work happens.
FIELD NOTES
USERS
Do: Start using Codex across full workflows as OpenAI is moving from helping you work to doing the work for you.
Don't: Assume current pricing or product boundaries will hold as consolidation and monetisation are already underway.
ANALYSTS
Do: Follow where usage is embedding into real workflows as that is where long-term value is being built.
Don't: Dismiss leadership and narrative pressure as noise as it is starting to shape policy, partnerships, and adoption.
DEVELOPERS/PRODUCTS
Do: Build inside agent-driven systems and execution layers, as that is where OpenAI is concentrating capability.
Don't: Overinvest in standalone features that can be absorbed into core products with one release cycle.
INVESTORS/VC
Do: Back companies that benefit from dependency on OpenAI’s execution layer rather than those competing alongside it.
Don't: Assume scale protects OpenAI as legal, regulatory, and narrative risks are rising at the same time as growth.
COMPANY ↓
OpenAI, staff/leadership profiles, board & governance, mission & values, HQ & global offices, funding rounds & investors, financial health & revenue, partnerships & alliances, organisational structure & hierarchy, regional presence & expansion.
OpenAI executive Kevin Weil, OpenAI’s chief technology officer of enterprise applications, Srinivas Narayanan and Bill Peebles, head of Sora, left the company. /OpenAI + WIRED + OpenAI + Tech Crunch
OpenAI’s President and Co-founder Greg Brockman announced that Stargate is a step towards meeting the demand of the compute-powered economy. /OpenAI
OpenAI acqui-hired Hiro, shutting down its AI finance product to fold the team into its broader roadmap. /OpenAI + Tech Crunch
OpenAI acqui-hired Cirrus Labs to strengthen agent engineering infrastructure, winding down its CI/CD products. /OpenAI
OpenAI rolled out ads in Australia, New Zealand, and Canada for users on Free and Go plans. /OpenAI
OpenAI investor SoftBank invited more banks to join its $40B loan backing its OpenAI investment. /Bloomberg
OpenAI announced its first permanent London office; days later, Anthropic unveiled plans for major UK expansion. /CNBC
OpenAI agreed to pay Cerebras $20B+ to use its server chips, double the amount previously associated with the deal. /The Information
TECHNOLOGY ↓
ChatGPT ecosystem, DALL·E tools, Whisper audio, product roadmap, prototypes, developer access, integrations, extensions, research, innovations
OpenAI released a major update to Codex, making it a more powerful partner for the more than 3 million developers. /OpenAI
OpenAI introduced GPT‑Rosalind for life sciences research. /OpenAI
OpenAI scaled up its Trusted Access for Cyber (TAC) program to verified individual defenders and teams responsible for defending critical software. /OpenAI
OpenAI expanded its cyber defence programme with funding, partnerships and controlled model access. /OpenAI
OpenAI brought GPT-5.4 and Codex into Cloudflare’s Agent Cloud, enabling enterprises to deploy AI agents at scale [VID]. /OpenAI + OpenAI
OpenAI added support for 90+ plugins in Codex, giving it more ways to gather context and take action across the tools users already use. /OpenAI + OpenAI
OpenAI launched the in-app browser inside Codex with comment mode. /OpenAI
To celebrate its 1st anniversary, OpenAI allowed Codex to reset its own rate limits across all plans. Enjoy all the new features. /OpenAI
OpenAI clarified a few things about Pro plan rate limits. /OpenAI
OpenAI responded to feedback by shipping updates shipped to Windows first, then Intel Mac, and Apple Silicon. /OpenAI
OpenAI announced an updated Agents SDK to help developers build agents that can inspect files, run commands, edit code, and work on long-horizon tasks within controlled sandbox environments. /OpenAI
OpenAI launched its Agents SDK with E2B as a sandbox partner for running persistent, long-running agents. /OpenAI
OpenAI research lead Joy Jiao and product lead Yunyun Wang joined
AndrewMayne on the OpenAI Podcast, episode 16, to discuss how OpenAI is building models for biology, drug discovery, and translational medicine [VID]. /OpenAI
OpenAI shared details of how users can create and refine videos, animations, charts, captions, and motion graphics with the @Remotion plugin in Codex. /OpenAI
OpenAI shared details of how users can inspect builds, understand failing jobs, and debug CI issues with the @CircleCI plugin in Codex. /OpenAI
OpenAI launched heartbeats in Codex: automations that maintain context inside a single thread over time. /OpenAI
OpenAI shared 3 ideas you can build with the macOS apps plugin in Codex. /OpenAI
OpenAI shared real-world Codex use cases across engineering and business workflows, signalling broader adoption beyond coding. /OpenAI
OpenAI partnered with JetBrains to host a Codex hackathon on 18-19 June, in San Franciso pushing deeper integration into developer IDE workflows. /OpenAI
OpenAI shared a video, ‘Bala Vs Waterfall, with ChatGPT’ [VID]. /OpenAI
OpenAI shared a video, ‘What Codex Unlocks for Wasmer’ [VID]. /OpenAI
STRATEGY/IMPACT ↓
Public Policy, CSR, ESG practices, ethics & AI safety, risk management, competitive strategy, thought leadership, industry impact, market positioning & competitive strategy
OpenAI’s Global Policy Chief, Chris Lehane, called out AI “doomers” and said, “When you put some of those thoughts and ideas out there, they do have consequences.” /San Francisco Standard
OpenAI’s Chief Revenue Officer, Denise Dresser, shared an internal memo about beating the competition. /The Verge
OpenAI released a new research report ‘Reigniting the discovery
engine for tomorrow’s cures’: why AI matters for the life sciences. /OpenAI + OpenAI
OpenAI highlighted a top safety ranking in TeenAegis’ AI Danger Index, reinforcing its child protection positioning. /OpenAI
Greg Brockman, President and Co-founder, shared his views on how the world is transitioning to a compute-powered economy. /OpenAI
MISCELLANEOUS ↓
Legal, scandals, industrial action, revenue, profiles, public policy, CSR, ESG, corporate culture, ethics, risk management, Sam Altman investments, +
OpenAI backed the extreme AI liability bill this week; Anthropic opposed it. /WIRED
OpenAI’s Chief Revenue Officer Denise Dresser shared what she’s learned since starting at OpenAI, 90 days in and 400+ customer meetings later [VID]. /OpenAI
OpenAI CEO Sam Altman’s house was targeted in a second attack; two suspects were arrested. /SF Standard
RESOURCES ↓
POVs, data visualisations & dashboards, key insights & metrics, industry reports & white papers, competitive analysis, recommended reading & book lists, analysis pieces & case studies +
Sam Altman’s side hustles blur the line between OpenAI’s interests and his own. /WSJ
Ronan Farrow on Sam Altman’s ‘unconstrained’ relationship with the truth [VID]. /The Verge
OpenAI’s oddly socialist, wildly hypocritical new economic agenda. /Vox
Mathematician Terence Tao on the GPT-5.4 pro solving Erdős problem #1196. /@haider1
Data centre delays threaten to choke AI expansion. /FT
OpenAI's policy doc mentions a public wealth fund, and Elon Musk touts universal income as a remedy to AI-driven unemployment. /Forbes
Not subscribed yet?
Join the thousands who are getting updated every week.
Try it for a month. Cancel anytime.
