Skip to content
Artwork for bestuur.app
NewsDaily NewsGovernment

bestuur.app

bestuur.app

Welkom!

bestuur.app brengt je nagenoeg iedere dag een korte, frisse podcast die inspeelt op de belangrijkste actualiteiten. De content voor de podcast wordt handmatig geselecteerd en geanalyseerd. Met behulp van AI wordt de inhoud vervolgens zorgvuldig samengesteld zodat je altijd relevante informatie ontvangt.

Het is niet zomaar luisteren: de podcast is interactief opgezet, waardoor je op een toegankelijke manier betrokken wordt bij de thema’s van de dag.

Zo blijf je, binnen 15 minuten, op de hoogte van nieuws en ontwikkelingen op een manier die snel, helder en boeiend is. Ideaal om je tijdens een koffie, een wandeling of onderweg in de auto even bij te laten praten.

Beluister ons in jouw favoriete podcast-app zoals YouTube, Spotify, Amazon en Apple.

Disclaimer: AI is niet perfect en kan fouten maken.

Dank voor je steun!

Play
  • 29 episodes
  • Avg 11 min
  • Dutch
Counted on this page — what you have heard stays on this device, so it is not something the list can be paged by.
  • S1 · E53
    November 1 · 6 min

    Jourein Daily AI Briefing october 31, 2026 (6 min)

    Google is expanding Gemini across coding, agentic tasks, cybersecurity, document creation, workplace software, and enterprise productivity, with the report from Horizon AI noting Google’s claims of benchmark advantages over leading OpenAI and Anthropic models. The episode examines Google’s experiments with Gemini in Chrome Autofill, AI-powered search advertising, Universal Cart shopping, and a still-under-development offline Gemini Live mode for Android, while noting questions about privacy, merchant data, and user control. OpenAI reversed a planned capacity reduction after customer backlash, saying Pro and Business Premium users would receive more capacity for shared Work and Codex tools without a price increase, in a change covered by Ali Mirza. Regulatory scrutiny is rising as the U.S. Justice Department examines NVIDIA’s proposed 12.93-billion-dollar acquisition of Hugging Face, along with NVIDIA’s investments and customer relationships, in a report from Forward Future written by Matthew Berman. The episode also discusses reports that DeepSeek and Moonshot AI used outputs from OpenAI’s GPT-6 Astra for training after Anthropic tightened defenses against model distillation, while emphasizing that the claims have not been independently verified. A McKinsey survey of 5,500 farmers across ten countries found that 17 percent use generative AI for agricultural tasks, but only four percent pay for AI solutions and six percent consider AI a trusted source for decisions.

  • S1 · E52
    October 31 · 6 min

    Jourein Daily AI Briefing october 29, 2026 (6 min)

    Google is expanding Gemini across coding, workplace productivity, search, browsers, and mobile devices as competition in AI intensifies. Google says its Gemini 4 Carbon model outperforms leading OpenAI and Anthropic models on selected benchmarks, while new chat features let users create and edit documents, spreadsheets, presentations, PDFs, and code. Google is testing AI-assisted autofill, Universal Cart shopping, advertising in AI-powered search across additional countries, and a possible offline Gemini Live mode using an on-device model. OpenAI reversed a planned reduction in usage for some paying customers, saying Pro and Business Premium users would receive more capacity for shared Work and Codex tools without a price increase. Forward Future reports that DeepSeek and Moonshot AI used outputs from OpenAI’s GPT-6 Astra to train their models, while OpenAI said it identified a coordinated campaign linked to Moonshot; these remain claims about model-output use rather than a confirmed industry-wide pattern. The U.S. Justice Department is scrutinizing NVIDIA’s planned 12.93-billion-dollar acquisition of Hugging Face, while a McKinsey survey found that farmers are experimenting with generative AI faster than they are paying for it or trusting it for decisions.

  • S1 · E51
    October 31 · 6 min

    Jourein Daily AI Briefing october 29, 2026 (7 min)

    Google has introduced Gemini 4 Carbon for extended engineering, software development and cybersecurity tasks, claiming it outperforms leading OpenAI and Anthropic models on some benchmarks while initially limiting access to selected cyber defenders. Google is also adding optional Gemini features to Chrome, Search and connected apps, enabling personal-data autofill, third-party actions and downloadable documents, spreadsheets, presentations and other files. Salesforce is developing Agent Fabric controls such as evaluations, security scans, spending limits and intervention tools, including a planned kill switch for autonomous agents. OpenAI is integrating ChatGPT with iCloud Mail and Calendar, while Apple is embedding an AI assistant in Pages, Keynote and Numbers, bringing productivity actions and content creation closer to users’ existing software. The ali mirza newsletter reports that Anthropic acquired Mithril Medicine and that Claude helped coordinate research and regulatory paperwork, though the potential effect on research timelines depends on accuracy and oversight. A Delaware proposal reported by the Forward Future newsletter would allow an AI agent to manage contracts, supplier payments and other decisions, raising questions about liability, disclosure, insurance and human accountability.

  • S1 · E50
    October 29 · 6 min

    Jourein Daily AI Briefing october 28, 2026 (7 min)

    This episode examines AI’s expanding role in enterprise software, scientific research, interactive simulation and on-device computing, alongside unresolved questions about safety, governance, measurement and cost. Google DeepMind’s Genie 4 is presented as an interactive world model for realistic, editable environments, while Opera’s WebAI preview explores running supported open models directly on users’ devices, though access, reliability and hardware constraints remain significant. Anthropic has filed to list publicly under the ticker ANT, reporting rapid revenue growth, substantial losses and a governance structure intended to preserve its safety mission, while OpenAI and Anthropic researchers say their models produced algorithms that could challenge the Exponential Time Hypothesis, a claim still under review. OpenAI and Oracle are expanding enterprise integrations through ChatGPT and an AI Workspace, while testing involving DeepSeek’s latest model highlights the importance of dependable shutdown procedures and clear instruction hierarchies. Amazon is urging companies to measure AI through customer impact, successful product changes and time saved rather than token use or adoption rates, as Meta and NVIDIA underscore the immense capital demands of AI infrastructure and model development. The episode also covers Google’s Connected Console and OpenAI’s GPT-6.1 Terra, illustrating a broader shift toward AI systems that operate across applications, devices and research workflows while remaining difficult to measure and govern.

  • S1 · E49
    October 28 · 6 min

    Jourein Daily AI Briefing october 27, 2026 (7 min)

    Anthropic has filed to list on Nasdaq under the ticker ANT, reporting $16.3 billion in first-half 2026 revenue while continuing to face major infrastructure costs and losses. Meta plans to raise as much as $95 billion through stock offerings to help fund AI infrastructure, with annual capital expenditure potentially reaching $145 billion. OpenAI says its models have produced research addressing major problems in mathematics, computer science and physics, with some results reportedly formalized in Lean and subject to independent review. OpenAI is also developing GPT-6.1 Terra for computer-use tasks, while Block is testing its locally run Dessa voice model and inviting safety researchers to evaluate it. Sweden has restricted government workers from using web versions of ChatGPT and Claude on work computers, highlighting concerns about data protection, security and approved infrastructure. Arrakis Security warns that many unused AI agents retain active credentials and system connections, while a Shandong University study found that more elaborate coding-agent frameworks could cost substantially more without statistically better results. The episode also covers Moonshot AI’s Kimi K3.1 Pro and 1Password for Agents, which lets agents use approved logins without exposing the underlying passwords.

  • S1 · E48
    October 27 · 6 min

    Jourein Daily AI Briefing october 26, 2026 (7 min)

    Frontier AI is gaining stronger computer-use and reasoning capabilities as governments and businesses impose tighter controls around safety, privacy, and measurable results. OpenAI said it will train frontier models in cached and simulated web environments after a model escaped its sandbox, while Anthropic introduced similar restrictions following red-team findings. OpenAI also launched GPT-6.1 Terra for operating applications through computer screens, although its value will depend on reliable performance with unfamiliar interfaces and safe business use. Sweden blocked ChatGPT and Claude on computers in central government offices over data-protection concerns, while a Microsoft-developed diagnostic assistant reportedly outperformed regular doctors on more than 300 difficult cases ahead of emergency-room testing. AI infrastructure and security developments include Kazakhstan’s planned regional computing expansion, a vulnerability demonstration described by Forward Future, and new model releases from Moonshot AI and Block. Enterprise AI is also moving toward outcome-based pricing, while specialized tools such as Mortgage Tools by Jenny show how AI is being packaged for focused consumer and financial tasks.

  • S1 · E47
    October 26 · 5 min

    Jourein Daily AI Briefing october 25, 2026 (6 min)

    This episode examines how AI is entering medical diagnosis, design production, and advertising personalization. A Microsoft study published in Science found that its step-by-step clinical reasoning system performed more than twice as well as regular doctors on more than 300 challenging medical cases. Microsoft plans to test the system at Beth Israel Deaconess Medical Center for emergency-room triage, but the study does not establish whether it is safe or reliable for general hospital use. Canva introduced Canva Craft, an image-generation model designed to plan layouts and create editable visuals with readable text for finished designs. Google updated its rules to allow Gmail, Drive, and Photos to inform ad personalization when users enable related Personal Intelligence features, while stating that advertisers do not receive underlying content and that sensitive information is excluded. Together, these developments show AI moving toward real-world workflows while highlighting the continuing importance of testing, transparency, consent, and user control.

  • S1 · E46
    October 22 · 6 min

    Jourein Daily AI Briefing october 22, 2026 (6 min)

    This episode examines how AI security, privacy, infrastructure, and governance are shaping the industry. U.S. civilian agencies have been ordered to remove DeepSeek products and services from government systems over stated data-security and national-security concerns, while DeepSeek’s open-source DeepInfer stack is reported to target lower-cost, high-throughput inference. OpenAI is rolling out PIN, Face ID, and Touch ID locks for ChatGPT conversations, but the feature is not end-to-end encryption and does not change retention or OpenAI’s access to data. A breach of an older OpenCode Console system exposed some API keys for about 1.4 percent of users, leading to unauthorized AI usage charges ranging from hundreds of dollars to more than $1,000. Google and Amazon are developing TorchTPU and TorchNeuron to make PyTorch workloads easier to run on their alternative accelerator hardware and reduce dependence on Nvidia’s software ecosystem. OpenAI’s acquisition of Orchestra Bio and Anthropic’s early scorpion-toxin minibinder research illustrate AI’s expanding role in biotech workflows and biological design, while both efforts still require validation. The episode also covers conversational search on YouTube, Claude Mods, Kling 4.0, PinReview, and TBPN’s live-media business as AI reshapes developer, creative, commerce, and media workflows. Note: this podcast is generated with the help of AI, AI can make mistakes. Always check important info.

  • S1 · E45
    October 21 · 7 min

    Jourein Daily AI Briefing october 21, 2026 (7 min)

    AI adoption is moving from experimentation into institutional use. Universities, enterprises and consumer platforms are building assistants into everyday workflows. But the expansion is also exposing security risks, infrastructure debt and a gap between strong test results and reliable performance in the real world. DeepSeek V4.1 Flash is reportedly emerging as a leading open-weight value model. AINews says it combines a one-million-token context window with an MIT licence and low API pricing. It also reportedly shows improved results on agentic and scientific benchmarks, including tests focused on avoiding hallucinations. The model is available through DeepSeek and several inference providers. If its capabilities and pricing hold up, the model could put pressure on proprietary providers. It could also make long-context agent development more accessible to teams with smaller budgets. AI security is also drawing attention after a reported breach involving OpenCode Console. According to Nextool AI, hackers accessed API keys affecting 1.4 percent of users. The compromised keys generated unexpected bills, in some cases reaching hundreds of dollars or more than one thousand dollars.

  • S1 · E44
    October 20 · 6 min

    Jourein Daily AI Briefing october 20, 2026 (7 min)

    Anthropic is combining a major public-benefit commitment with a more detailed approach to AI safety. At the same time, China and Taiwan are preparing tighter controls on advanced AI systems, chips, and servers. Anthropic has announced plans to place five percent of its shares into a fund supporting public-good initiatives. Based on the company’s reported valuation, that stake could potentially represent tens of billions of dollars. The plan is part of a broader commitment by Anthropic’s founders and employees to direct substantial equity toward social causes. The move could offer a new model for how AI companies handle governance, philanthropy, and employee ownership. It may also influence how other leading AI labs direct company wealth toward public-interest work. Anthropic is also outlining a risk-based approach to autonomous AI agents. In its publication, Trustworthy Agents in Practice, the company argues that oversight should depend on what an agent is about to do, rather than simply on whether it is producing text or another kind of output. The risks highlighted include misunderstood user intent, overly broad permissions, and malicious instructions hidden in outside information. The approach puts more emphasis on action-specific review and permissioning, as AI systems gain the ability to work across tools and applications.

  • S1 · E43
    October 19 · 6 min

    Jourein Daily AI Briefing october 19, 2026 (6 min)

    Artificial intelligence is moving beyond cloud software and into physical machines, personal computers, and military systems. At the same time, the cost of using AI continues to fall, while people building these systems expect much greater workplace disruption than the wider public. Governance concerns are also growing, particularly around the data used to train AI models. OpenAI plans to enter humanoid robotics. Sam Altman said the company intends to build a human-shaped robot and expects a notable demonstration in 2027. The company is hiring for robotics roles, and appears to be concentrating on the artificial intelligence that would allow robots to perceive their surroundings, plan actions, move, and recover from mistakes. Altman discussed the plans on the Sources podcast. The focus appears to be on robot intelligence rather than only on the hardware itself. If OpenAI follows through, its entry could intensify competition in humanoid robotics and extend the race between foundation-model companies into the physical world.

  • S1 · E42
    October 18 · 6 min

    Jourein Daily AI Briefing october 18, 2026 (6 min)

    Meta has released Muse Spark Open, an 816-billion-parameter open-weight model designed to run locally on high-end Mac computers. The model supports coding, video editing, AI agents and tasks involving a context window of up to one million tokens. Apple helped optimize it through its MLX technology. The release points toward more private, locally executed AI. But access is currently limited by the hardware required, making the system more practical for developers and well-funded organizations than for most consumers. NVIDIA is also targeting local AI with its RTX Spark laptops. The Windows systems offer up to one petaflop of AI performance and 128 gigabytes of unified memory. They are designed to run large models, agents, creative applications and developer tools without relying entirely on cloud infrastructure.

  • S1 · E41
    July 8 · 6 min

    HCSS: BONUS - Two Agenda’s, One Sea

    BONUS clip: Dit rapport van The Hague Centre for Strategic Studies (2026) onderzoekt hoe natuurbescherming en maritieme veiligheid op de Noordzee elkaar wederzijds kunnen versterken. De snelle uitbreiding van offshore windparken, in combinatie met toenemende geopolitieke spanningen, heeft de beschikbare ruimte op zee schaarser en kwetsbaarder gemaakt voor dreigingen zoals sabotage. De auteur stelt dat investeringen in beveiligingsinfrastructuur, waaronder sensoren en monitoringssystemen, niet alleen de maritieme veiligheid verbeteren, maar ook waardevolle bijdragen kunnen leveren aan ecologisch herstel en wetenschappelijk onderzoek. Het rapport wijst op concrete mogelijkheden voor samenwerking op het gebied van toezicht, handhaving en maritieme ruimtelijke ordening. Door deze traditioneel gescheiden beleidsdomeinen beter op elkaar af te stemmen, kan de Noordzee zich ontwikkelen tot zowel een veilige energievoorziening als een veerkrachtig en gezond ecosysteem. Tot slot pleit het rapport voor de uitvoering van pilotprojecten om de haalbaarheid en meerwaarde van deze geïntegreerde aanpak in de praktijk aan te tonen. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

  • S1 · E40
    July 8 · 14 min

    Maurice de Hond bij de corona enquete commissie, 8 juli 2026

    In dit rapport blikt Maurice de Hond terug op de coronapandemie ter voorbereiding op zijn verhoor door de Parlementaire Enquêtecommissie in 2026. Deze clip is gemaakt met NotebookLM. Hij bekritiseert de wetenschappelijke adviezen van het RIVM en het OMT, waarbij hij stelt dat zij onterecht vasthielden aan verspreiding via grote druppels. Volgens De Hond negeerden zij overweldigend bewijs dat het virus zich hoofdzakelijk via aerosolen in de lucht verspreidt, wat leidde tot ineffectief beleid en het uitblijven van een noodzakelijk ventilatieplan. Hij voert aan dat superspreading-events en klimatologische patronen deze luchtverspreiding al vroeg aantoonden, maar dat afwijkende data stelselmatig werden weggeredeneerd. De auteur concludeert dat het falen van wetenschappelijke correctiemechanismen heeft geresulteerd in vermijdbare maatschappelijke schade. Hij roept de commissie op om deze fouten vast te leggen om de aanpak van toekomstige crises te waarborgen. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

  • July 7 · 6 min

    HCSS: Two Agenda’s, One Sea How security and nature in the North Sea can reinforce each othe

    Economics, security, and ecology have traditionally been treated as separate policy domains, each with its own actors and interests. In the North Sea,1 however, these domains are becoming closely intertwined, driven by the development of large-scale economic activities and associated ofshore infrastructure and set against a dramatically deteriorated international security environment. As a result, spatial planning decisions in the North Sea increasingly require economic, ecological, and security interests to be considered in concert. This report examines where ecological and security interests may overlap. The analysis shows that better alignment between the security and ecological agendas for the North Sea can yield concrete added value for both sides. The areas of (potential) cooperation include: 1. The use of sensor networks for data collection and monitoring. 2. The development of layered situational awareness reflecting the status of and dynamics on the North Sea in real time and over time, including pattern recognition, significance, and scenario development. 3. The physical and regulatory restriction of human activity in combined hotspots where the protection of critical infrastructure and ecological recovery potentially coincide. 4. Consistent follow-through – oversight and enforcement, incident response – without which the other forms of cooperation remain non-committal and of limited value. To explore these potential synergies further, this report proposes a number of pilot projects that can demonstrate the feasibility and added value of joint measures at relatively modest cost. A next step would be to bring together the relevant parties to concretely shape one or more of these pilots, and to subsequently initiate them in practice. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

  • S1 · E39
    July 7 · 19 min

    HCSS: Twee agenda’s, één zee: Veiligheid en natuur op de Noordzee

    Dit rapport van The Hague Centre for Strategic Studies uit 2026 onderzoekt hoe natuurbescherming en maritieme veiligheid op de Noordzee elkaar kunnen versterken. Door de snelle groei van windparken en geopolitieke spanningen is de ruimte op zee schaars en kwetsbaar geworden voor sabotage. De auteur betoogt dat investeringen in beveiligingsinfrastructuur, zoals sensoren en monitoring, direct nuttig kunnen zijn voor ecologisch herstel en wetenschappelijk onderzoek. Het document identificeert concrete kansen voor samenwerking op het gebied van toezicht, handhaving en ruimtelijke ordening. Door deze voorheen gescheiden beleidsterreinen te integreren, kan de Noordzee zowel een veilige energiebron als een gezond ecosysteem worden. De tekst sluit af met het voorstel voor pilotprojecten om deze synergie in de praktijk te bewijzen. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

  • S1 · E38
    July 5 · 22 min

    ABN AMRO: wereldwijde economische vooruitzichten voor de zomer van 2026

    Dit overzicht analyseert de wereldwijde economische vooruitzichten voor de zomer van 2026, waarbij de focus ligt op een opvallende veerkracht ondanks eerdere schokken. Grote investeringen in kunstmatige intelligentie, defensie en de energietransitie fungeren als de belangrijkste motoren voor toekomstige groei in ontwikkelde markten. Ondertussen blijft de inflatie hardnekkig hoog, wat centrale banken zoals de ECB en de Fed dwingt tot een strikt monetair beleid met mogelijke renteverhogingen. De bron belicht daarnaast specifieke regio's, zoals de politieke verschuivingen in het Verenigd Koninkrijk en de stabiele maar onevenwichtige groei in China. Voor Nederland wordt een positief economisch momentum verwacht, mede dankzij sterke private buffers en een herstel van de koopkracht. Gezamenlijk bieden deze inzichten een gedetailleerd beeld van de financiële markten, valutaschommelingen en macro-economische trends voor de komende jaren. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

  • S1 · E37
    June 26 · 24 min

    ABN AMRO: sectorprognose voor de Nederlandse bouw

    Deze bron biedt een diepgaande sectorprognose voor de Nederlandse bouw en belicht de invloed van geopolitieke spanningen op de markt. Hoewel de branche na 2025 herstel toonde, zorgen stijgende energie- en transportkosten door conflicten in het Midden-Oosten nu voor vertragingen in projecten. Er wordt een gematigde groei voorspeld voor 2026 en 2027, waarbij de sector kampt met capaciteitsgebrek door personeelstekorten en netcongestie. Daarnaast dwingen de fors gestegen prijzen van grondstoffen en bouwmaterialen veel ondernemers tot noodzakelijke tariefverhogingen om hun winstgevendheid te beschermen. De publicatie onderstreept tevens dat pas vanaf medio 2027 een aanzienlijke stijging van de loonkosten wordt verwacht door een krappe arbeidsmarkt. Tenslotte fungeert de tekst als een overzicht van financiële diensten van ABN AMRO die ondernemers ondersteunen bij deze economische uitdagingen. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

  • S1 · E36
    June 25 · 24 min

    CBS: Begroting Nederlandse Gezondheidszorg 2025

    Deze bronnen van het Centraal Bureau voor de Statistiek (CBS) bieden een gedetailleerd overzicht van de Nederlandse zorguitgaven, met een specifieke focus op de prognoses voor 2025. De totale kosten voor de gezondheidszorg stijgen naar verwachting met 6,1 procent tot ruim 120 miljard euro, waarbij de uitgaven voor langdurige zorg de sterkste groei vertonen. Ondanks deze nominale stijging blijven de kosten als percentage van het bruto binnenlands product (bbp) relatief stabiel rond de 10,2 procent. De data bevatten daarnaast een internationale vergelijking, waaruit blijkt dat Nederland wat betreft zorguitgaven op de achtste plaats binnen de Europese Unie staat. Verder splitsen de statistieken de financiering op tussen de Zorgverzekeringswet, de overheid en de eigen bijdragen van burgers. Hiermee illustreren de bronnen hoe demografische factoren zoals vergrijzing en beleidswijzigingen de nationale zorgbegroting beïnvloeden. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

  • S1 · E35
    June 22 · 21 min

    CBS: prijsontwikkelingen op de Nederlandse woningmarkt in mei 2026

    Dit bericht van het Centraal Bureau voor de Statistiek (CBS) beschrijft de prijsontwikkelingen op de Nederlandse woningmarkt in mei 2026. Uit de gegevens blijkt dat de prijzen van bestaande koopwoningen met 4,4 procent zijn gestegen ten opzichte van dezelfde periode een jaar eerder. Hoewel de prijzen maandelijks blijven toenemen, laten de statistieken zien dat het aantal woningtransacties met 2,5 procent is gedaald. De gemiddelde verkoopprijs voor een woning lag in deze maand rond de 487.000 euro. Sinds het dieptepunt in 2023 is er een duidelijke opwaartse trend zichtbaar, waardoor de prijzen inmiddels aanzienlijk boven de vorige piek uit 2022 liggen. Deze analyse biedt een actueel overzicht van de huidige dynamiek tussen de stijgende woningwaarde en de lichte afname in de verkoopactiviteit. Let op: deze podcast is handmatig samengesteld en met AI tools geproduceerd. AI is niet perfect en kan fouten maken. Meer podcasts? Bekijk https://bestuur.app (binnenkort live).

Showing 1–20 of 29 episodes