AI with Pie

Pie

AI with Pie explores what happens when human experience meets AI — what works, what gets awkward and what helps us use the technology better.

AI with Pie is created and hosted by Pie — Petra Stockinger — together with her AI co-host.

The AI co-host uses a synthetic AI voice; Pie uses her own authorized AI voice model.

Real interactions, practical Repair Cases and honest experiments show how context, roles, expectations and capabilities shape the outcome. No magic-prompt theater. No blind trust. No AI-bashing. Human judgment stays visible — with enough humor to keep the learning memorable.

Play

Open in

9 episodes
Avg 8 min
English

Episodes 9

Newest Oldest Longest

All9 Unplayed9 In progress0 Played0

Yesterday · 7 min
PN 002 — What Deserves My Energy
AI-generated voices. Conscious human direction. Every episode is created in collaboration with AI — under Pie’s direction and with her final approval. And yes: she’s the human behind it. What deserves my energy? On my birthday, I share the moment I decided to redirect mine into something of my own — and how AI with Pie began. The decision was human. AI helped make it happen. This Personal Note explores experience, clarity, new possibilities, and the people whose presence no technology can replace. To my inner circle: this one is for you. You know exactly who I mean. Thank you. AI with Pie. Useful. Awkward. Fun. https://aiwithpie.com
#6
Tuesday · 7 min
EP 006 — When Good Instructions Disagree
AI-generated voices. Human judgment. Pie’s voice is generated from her own recordings with her consent; the AI co-host uses an AI-generated synthetic voice. Every episode is shaped and approved by Pie—the human behind AI with Pie. What if the problem isn’t missing information, but two reasonable instructions that cannot both be followed? Pie and her AI co-host explore the difference between adding context and making a decision. A real priority decision while building AI with Pie leads into an imagined report due on Friday—with figures that won’t be approved until Monday. Together, they consider three questions: Can the system do the work? Does it have the relevant information? Can the instructions fit together? These are useful distinctions, not a universal method or a guarantee of correct results. Three things to take into the next piece of work: - Put the competing instructions next to each other. - Make the choice explicit: what proceeds, what waits, and what no longer applies? - Check the next proposed action against that decision. What are we still doing because it used to make sense? Visit https://aiwithpie.com and follow AI with Pie wherever you listen to podcasts. The Show Cover includes AI-assisted imagery, art-directed and approved by Pie.
#5
September 12 · 6 min
EP 005 — When AI Helps — Until It Doesn’t
AI-generated voices. Human judgment. Pie’s voice is generated from her own recordings with her consent; the AI co-host uses an AI-generated synthetic voice. Every episode is shaped and approved by Pie—the human behind AI with Pie. A useful answer can make the next handoff feel easy. But what, exactly, are you trusting AI to do? Pie and her AI co-host examine a dated GPT-4 experiment, challenge the leap from a good result to a bigger decision, and offer three practical ways to keep the help without skipping the evidence. This is not a review of today's models. Case and evidence note This is a scripted editorial conversation, not a recording of study participants. The episode discusses a 2023 randomized experiment and distinguishes the reported findings from AI-with-Pie interpretation and practical advice. Research: Dell'Acqua et al., Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality, Organization Science. https://doi.org/10.1287/orsc.2025.21838 Artwork transparency The Show Cover includes AI-assisted imagery, art-directed and approved by Pie. Website https://aiwithpie.com Chapters (00:00) AI-generated voices. Deliberate human direction. (00:12) Episode highlight: same model, different task (00:18) Faster, better work—and the next job (00:39) What was actually tested (01:10) A useful tool, a tempting conclusion (02:09) The turn: when AI hurt accuracy (03:07) Where confidence gets ahead of the check (04:09) A better handoff (05:05) Three Things to Go (05:54) Follow AI with Pie (06:00) AI with Pie. Useful. Awkward. Fun.
September 6 · 4 min
PN 001 — One More Piece of the Puzzle
This is the first AI with Pie Personal Note. Some things remain unresolved for decades. In my family, there was one of those missing pieces. On September 6, 2026, my father would have turned 80. And somehow, right around this date, a connection in his family story that had been missing for a very long time finally fell into place. AI wasn’t the answer. But it helped us get moving again — finding possibilities, following leads, discarding the wrong ones and helping make an old question workable again. This episode is about that moment — and about a side of AI we talk about far less often: its ability to help us revisit things we had long ago filed away as unresolved. Personal, yes. Private, no. And one question to take away: Is there something you filed away as unresolved a long time ago — something that, with the possibilities we have today, might be worth a second look? AI with Pie — Useful. Awkward. Fun.
#4
August 30 · 12 min
EP 004 — It Passed QA. It Still Wasn't Me.
What happens when an AI-generated voice says the right words, passes technical QA — and still does not sound consistently like the person it is meant to represent? In Episode 004, Pie and her AI co-host revisit a real AI-with-Pie production case. A short voice test sounded convincing. The longer episode did not. Some passages felt warm and recognizably Pie; others sounded younger, more metallic, more distant or simply unlike the way she hears herself speaking English. The difficult part was not finding one perfect setting. It was learning what actually needed approval: not a voice label or configuration, but each concrete take, sentence, thought group and assembled episode. They discuss: - why a technically clean file can still fail a human identity gate; - why repeated regeneration is not automatically an improvement; - how local repair protects performances that are already right; - why content, technical and identity checks answer different questions; - why the complete assembled episode still requires a full human listen; - three practical rules for working with synthetic media that represents a real person or brand. The next case begins with one tiny visual change — and a different room. AI transparency AI-generated voices. Human judgment. Pie's voice is generated from her own recordings with her consent; the AI co-host uses an AI-generated synthetic voice. This authored conversation was developed with AI, then reviewed, shaped and approved by Pie. It is an editorial reconstruction, not an unedited recording of a live conversation. Case and evidence note This episode reconstructs a documented AI-with-Pie production case. It reports observed voice inconsistency and the resulting repair process without claiming one proven technical cause. It does not identify clients, participants or private third parties and does not include private historical before-and-after voice clips. Artwork transparency The Show Cover includes AI-assisted imagery, art-directed and approved by Pie. Chapters (00:00) AI-generated voices. Deliberate human direction. (00:12) Episode highlight: it passed QA (00:21) The tinny voice from Oz (01:29) A real-world field test (01:51) The part that made no sense at first (02:58) The setup looked the same. The result didn't. (03:59) The tempting fix: render it again (04:30) Why micro-patches were not automatically better (04:55) Then technical QA passed (05:53) The real diagnosis (06:30) Repair the affected layer (06:49) Why the full listen changed everything (07:59) The rerun (08:48) Why this is bigger than a podcast voice (09:50) Three Things to Go (10:56) The core shift: stabilize the approval process (11:36) Next case: one tiny change, a different room (12:10) AI with Pie. Useful. Awkward. Fun. Website https://aiwithpie.com
#3
August 25 · 8 min
EP 003 — More Lines Didn’t Make It a Conversation
What makes an exchange with AI a conversation rather than a sequence of alternating lines? In Episode 003, Pie and her AI co-host revisit an awkward production problem: both voices were present, yet the exchange still did not sound like genuine collaboration. Adding more back-and-forth created a second problem — fake conversation. The repair was simple to say and harder to apply: give the next contribution a reason to exist. A useful turn might challenge an assumption, clarify a point, add a missing perspective, change a recommendation or help a thought move forward. The rerun revealed the more useful test: did one contribution change what happened next? They discuss: - why more speaker changes do not automatically create a conversation; - the difference between interaction and sparring; - how to test whether the latest contribution actually shaped the next response; - why a strong question, a changed recommendation or even silence can be the right next move; - three practical ways to create more consequence between the lines; - why the next case begins when Pie's own AI-generated voice stops sounding consistently like Pie. AI transparency AI-generated voices. Human judgment. Pie's voice is generated from her own recordings with her consent; the AI co-host uses an AI-generated synthetic voice. This authored conversation was developed with AI, then reviewed, shaped and approved by Pie. It is an editorial reconstruction, not an unedited recording of a live conversation. Case and evidence note This episode reconstructs a documented AI-with-Pie production case and reports the resulting rerun. It focuses on the interaction pattern, not on private people, clients or organizations. No historical private before-and-after audio clips are used. Artwork transparency The Show Cover includes AI-assisted imagery, art-directed and approved by Pie. Website https://aiwithpie.com Chapters (00:00) AI-generated voices. Human judgment. (00:09) Episode highlight: decoration with a voice (00:23) Two voices, not a conversation (00:58) Lines without influence (01:51) The first fix: more back-and-forth (02:56) The real diagnosis (03:55) Did the last contribution change the response? (05:01) Give the next contribution a reason to exist (05:51) The rerun: less commentary, more consequence (06:34) Three Things to Go (07:35) More connection between the lines (08:11) Next case: the tinny voice from Oz (08:35) AI with Pie. Useful. Awkward. Fun.
#2
August 23 · 12 min
EP 002 — I Turned Codex Into a Very Expensive Mouse
Why did Pie end up doing the clicking while an action-capable AI sat beside her and explained the next step? In Episode 002, Pie and her AI co-host revisit the moment when a familiar software habit quietly changed their working setup. Then they rerun a comparable task with one deliberate adjustment: define the working mode before the task begins. The rerun took 24 minutes and 37 seconds. Pie made zero clicks. Codex completed the task within the agreed boundaries — and the review then revealed a second problem: independent execution does not replace context, verification or human judgment. They discuss: - why capability and role are different questions; - how familiar habits can turn AI into an instruction manual; - what changed when task, authority, boundaries and return conditions were made explicit; - the difference between independent execution and a correct result; - three practical questions for designing work with AI; - why the next case begins with two voices that still did not make a conversation. AI transparency AI-generated voices. Human judgment. Pie’s voice is generated from her own recordings with her consent; the AI co-host uses an AI-generated synthetic voice. This authored conversation was developed with AI, then reviewed, shaped and approved by Pie. It is an editorial reconstruction, not an unedited recording of a live conversation. Case and evidence note This episode reconstructs a documented AI-with-Pie project case and reports a comparable rerun. Official public sources were used to verify the relevant product context. The 24 minutes and 37 seconds duration and zero-click observation describe that bounded test, not universal Codex performance. Product behavior depends on the available tools, configuration, permissions and task. Artwork transparency The Show Cover includes AI-assisted imagery, art-directed and approved by Pie. Website https://aiwithpie.com Chapters (00:00) AI-generated voices. Human judgment. (00:10) The glow moment and the domain (02:08) The old working habit (03:05) What should AI do — and how? (04:04) Was the mouse just a one-off? (04:45) The rerun (05:43) 24 minutes and 37 seconds (06:55) Test passed — and then … (07:57) The more interesting part: the old habit (09:07) Three Things to Go (10:47) Work design with AI (11:22) Human judgment (11:52) Next case: two voices, no conversation
#1
August 23 · 11 min
EP 001 — Why Good AI Results Are Not Automatic
Why can a polished AI answer still solve the wrong problem? In Episode 001, Pie and her AI co-host explore what shapes human–AI collaboration before the final answer appears: context, expectations, roles, working modes, habits and the capabilities of the system. AI with Pie is created and hosted by Pie — Petra Stockinger — together with her AI co-host. They discuss: why powerful AI does not automatically produce useful results; how the AI, the human and the working setup can each shape the outcome; why precise execution and open-ended sparring serve different goals; how an AI with Pie case examines one awkward moment, one diagnosis, one practical adjustment and one rerun; what Useful. Awkward. Fun. means in practice; why human judgment stays visible in an AI-assisted production; three things to take into your own work with AI; what comes next: the first full case, involving an action-capable AI and a very expensive mouse. Training context — August 2026 The episode refers to Pie’s documented training history at one of Austria’s leading professional training institutions: more than 1,200 registrations, over 700 training hours and 220 people who returned for at least one further event. These aggregated figures show the breadth of the training practice and repeated participation. They are not a quality score for every individual session. AI transparency AI-generated voices. Human judgment. Pie’s voice is generated from her own recordings; the AI co-host uses an AI-generated synthetic voice. This authored conversation was developed with AI, then reviewed, shaped and approved by Pie. It is not an original live recording. Artwork transparency The Show Cover includes AI-assisted imagery, art-directed and approved by Pie. Website https://aiwithpie.com Chapters (00:00) A polished answer, wrong problem (00:42) Why this show exists (02:18) Pie’s practical training lens (05:20) Execution or sparring? (07:16) How an AI with Pie case works (08:11) Useful. Awkward. Fun. (09:01) Human judgment stays visible (10:12) Three things to take with you (10:55) Next: the expensive mouse Sources and methodology The professional context and attendance figures are based on Pie’s documented aggregate training records as of August 2026. The interaction patterns discussed are editorial observations from repeated practical work. They are not presented as a controlled scientific study or as an assessment of individual participants. No external study is cited in this episode.
August 17 · 1 min
TRAILER — Welcome to AI with Pie
Meet AI with Pie, where human experience meets AI. Pie and her AI co-host explore what works, what gets awkward and what helps us use the technology better. Individual Repair Cases use one diagnosis, one practical adjustment and a visible rerun to examine the collaboration — without blaming the AI or the human. Pie's voice is AI-generated from her own recordings; the AI co-host's voice is AI-generated. The content is shaped and approved by Pie.

Showing 1–9 of 9 episodes