Artificial intelligence dominated global headlines today as Jack Dorsey announced sweeping job cuts at Block Inc., the Mobile World Congress opened with AI as its undisputed star, and new analysis exposed persistent weaknesses in even the most advanced models when faced with everyday documents.
In an internal memo released Wednesday, Jack Dorsey, co-founder of Twitter (now X) and CEO of Block, confirmed the elimination of more than 4,000 positions — nearly half the company’s global workforce. The layoffs at the fintech giant behind Square, Cash App and its Bitcoin initiatives were not driven by financial distress but by AI’s ability to transform, accelerate and, in many cases, replace human work.
“AI does not reduce work. It transforms it, accelerates it, and in the most honest cases, replaces it,” Dorsey wrote. He added that most companies are “late” to this realization and predicted that within the next year the majority will undertake similar structural changes. “I’d rather get there honestly and on my own terms than be forced to do so reactively,” he stated. Block’s shares surged 23 % in after-hours trading following the announcement.
The move fits a broader pattern. The article cited Amazon’s 30,000 layoffs over the past five months, Pinterest’s 15 % workforce reduction in January, Salesforce’s cut of support staff from 9,000 to 5,000, and Duolingo’s termination of contracts with 10 % of its contractors — all explicitly linked to AI efficiencies. In 2025 alone, companies attributed 55,000 dismissals to AI, more than 12 times the figure from two years earlier.
Just hours after Dorsey’s memo, the 20th edition of the Mobile World Congress (MWC) opened its doors in Barcelona, drawing an expected 109,000 professionals, 2,900 exhibitors and around 60 government ministers. Organized by the GSMA, the four-day event has placed “AI sovereignty” and the total transformation of telecommunications at the center of discussions.
Telecom infrastructure is now seen as the essential backbone for AI’s expansion, enabling the fast, reliable and secure movement of massive data volumes. Sessions will address the rollout of 5G, the foundations of 6G, satellite connectivity and Europe’s push for greater digital autonomy. Major players including Samsung, Huawei, Nokia, Orange, Xiaomi, Honor, Google, Microsoft, Meta and Amazon are showcasing AI-integrated solutions, while SpaceX President Gwynne Shotwell is among the keynote speakers. Apple, as usual, is absent.
Honor is expected to steal attention with the world’s first “robot phone,” signaling the convergence of AI, mobility and robotics. Analysts describe the smartphone market’s surprising resilience thanks to aggressive Chinese manufacturers and fresh product launches.
Yet while business leaders and telcos bet heavily on AI’s potential, a detailed report from Xataka underscored its current shortcomings. Advanced models such as Google’s Gemini and OpenAI’s ChatGPT excel at solving complex mathematical equations and generating or reviewing code — feats that once seemed impossible. Programmers now routinely rely on tools like Claude for signing off on GitHub commits.
However, the same systems frequently collapse when asked to handle PDF files — a ubiquitous format in business, law and academia. Users report chaotic outputs: columns mixed together, footnotes inserted mid-sentence, tables rendered illegible and entire summaries that miss the document’s logical structure.
The root cause, the article explains, lies in the PDF format itself. Unlike HTML, which carries explicit semantic hierarchy (titles, paragraphs, tables), a PDF is primarily a visual description: text exists as independent fragments placed at precise coordinates on a page, with no inherent relationships between sentences or sections. Optical character recognition can extract words, but reconstructing the intended reading order and logical flow remains a major challenge.
This contrast — superhuman performance on structured tasks versus repeated failure on everyday unstructured documents — serves as a sobering reminder of AI’s uneven progress as the technology reshapes entire industries.
Together, today’s developments paint a complex picture: AI is already forcing painful but proactive corporate restructuring, powering the next generation of telecom infrastructure, and yet still stumbling over the practical realities of the documents that power the modern world. As Dorsey warned and MWC attendees will debate over the coming days, the AI revolution is here — transformative, disruptive, and still very much a work in progress.
Discussion about this post