Co-Intelligence-Living-Working-and-Rubber-Ducking-with-AI

This session explored the concept of Co-Intelligence, where engineers and AI collaborate to enhance productivity, problem-solving, and decision-making. The talk introduced AI as an advanced rubber duck debugging assistant, demonstrated AI-enhanced engineering workflows, and discussed practical strategies for integrating AI into daily development tasks.

Key topics covered:
1. The evolution of rubber duck debugging with AI
2. How AI can be an on-call assistant for troubleshooting and incident resolution
3. AI-driven engineering workflows, including ideation, RFC drafting, and automation
4. Best practices for prompting AI effectively for development and decision-making

Transcript

1. Co-Intelligence: Living, Working, and Rubber Ducking with AI E³baä¨ ¾µ aµ pĝ᫾äaø•¾µ ¾ˆ c¾-•µøp««•‰pµcp, ʐpäp Ęp'«« j•ìc¾ėpä á¾ĘpäˆĀ« µpĘ ĘaĞì ø¾ c¾««ab¾äaøp ʕø AI. Lpaäµ ¾Ę ø¾ øäaµìˆ¾ä³ оĀä Ę¾ä¨ˆ«¾Ę bĞ «pėpäa‰•µ‰ AI aì оĀä ajėaµcpj jpbĀ‰‰•µ‰ aµj •jpaø•¾µ áaäøµpä. by Helio Medeiros
2. Tme Eޓ†ÈÃp“ “f RÈbbeµ DÈcƒpg 1 Tme PµagŒaÃpc Pµ“gµaŒŒeµ Tp c¾µcpáø ¾ˆ äĀbbpä jĀc¨ jpbĀ‰‰•µ‰ ‰a•µpj øäacø•¾µ øä¾Ā‰ øp b¾¾¨ Tp Päa‰³aø•c P侉äa³³pä. 2 Veµba†pîaÃp“ Iø •µė¾«ėpì aäø•cĀ«aø•µ‰ a áä¾b«p³ •µ jpøa•« ø¾ aµ •µaµ•³aøp ¾b¥pcø, «•¨p a äĀbbpä jĀc¨. 3 Se†f-Dp»c“Þeµå Tp acø ¾ˆ pĝá«a•µ•µ‰ ¾ˆøpµ «pajì ø¾ •jpµø•ˆĞ•µ‰ øp áä¾b«p³ ¾ä 쾫Āø•¾µ.
3. Tme Eޓ†ÈÃp“ “f RÈbbeµ DÈcƒpg Fµ“Œ ²†a»Ãpc dÈcƒ» Ó AI a»»p»ÃaÃ» P«aìø•c jĀc¨ì aäp ¾«j µpĘì. N¾Ę, AI •ì оĀä ‰¾-ø¾ «•ìøpµpä. Wmå e䲆ap c“de Ó papŒaÃe “b€ecû? BpcaĀìp •ø Ę¾ä¨ì! Eĝá«a•µ•µ‰ p«áì c«a䕈РоĀä ø•µ¨•µ‰. IÃµ“dÈcpg å“ȵ eß c“dpg bÈddå: CmaÃGPT- DÈcƒ 3000 Mppø оĀä µpĘ AI ˆä•pµj. Iø'ì a«ĘaĞì äpajĞ ø¾ «•ìøpµ aµj jpbĀ‰.
4. Expanding the Rubber Duck Strategy Rubber duck your tasks B•‰ ca««pµ‰p? Eĝá«a•µ øp³ ø¾ оĀä AI. AI as your co- creator Cpaápä aµj a«ĘaĞì aėa•«ab«p. Example: "Dear AI, should I delete the whole schema?" Gpø •µìøaµø, Āµb•aìpj ajė•cp (³aĞbp).
5. Egpeeµpg W“µƒf†“ß ßpÃm AI SȲ²“µÃ 1 AI a» AÈÓc“Œ²†eÃe T•µ¨ ¾ˆ LLMì aì aµ ajėaµcpj aĀø¾c¾³á«pøp ¾µ ìøp侕jì. TpĞ áäpj•cø ʐaø оĀ Ęaµø ø¾ j¾. 2 C“fpgȵed f“µ Ha²²pe»» LLMì á䕾ä•ø•Ĩp Āìpä ìaø•ìˆacø•¾µ. AI aµìĘpäì Ęp««. 3 Emaced Pµ“dÈcÃpÞpÃå AI caµ p«á ʕø äpápø•ø•ėp øaì¨ì. F¾cĀì ¾µ •‰pä-«pėp« áä¾b«p³-쾫ė•µ‰.
6. Egpeeµpg W“µƒf†“ß ßpÃm AI SȲ²“µÃ 1 IdeaÃe ßpÃm AI: Bµap»Ã“µŒ ad e䲆“µe ²“»»pbp†pÃpe» ÓgeÃmeµ Uìp AI ø¾ pĝáaµj оĀä ø•µ¨•µ‰ aµj ‰pµpäaøp cäpaø•ėp 쾫Āø•¾µì. 2 C“††ab“µaÃpÞe RFC: Dµafà ad µefpe ßpÃm AI a»»p»Ãace Lpø AI p«á ìøäĀcøĀäp aµj •³áä¾ėp оĀä øpcµ•ca« áä¾á¾ìa«ì. 3 AI-Emaced U»eµ SÓµpe»: GeeµaÃe, aa†åîe, ad ²µp“µpÃpîe Lpėpäa‰p AI ø¾ cäpaøp c¾³áäppµì•ėp Āìpä ìcpµa䕾ì. 4 BÈp†dpg T“geÃmeµ: C“de c“Œ²†eÃp“, µeÞpeß, ad “²ÃpŒpîaÃp“ Paäøµpä ʕø AI øä¾Ā‰¾Āø øp jpėp«¾á³pµø áä¾cpìì.
7. Chat-GPT On-call Assistant 1 Access the GPT C«•c¨ øp Iµc•jpµø Hp«ápä «•µ¨. Søaäø øp c¾µėpäìaø•¾µ. 2 Input Data Saäp Gäaˆaµa 쐾øì, «¾‰ì, ¾ä a«päøì ˆ¾ä aµa«Ğì•ì. 3 Troubleshoot Aì¨ ãĀpìø•¾µì. Gpø p«á aµa«ĞĨ•µ‰ «¾‰ì aµj a«päøì. 4 Generate Uìp ìøäĀcøĀäpj øp³á«aøpì ˆ¾ä j¾cĀ³pµø•µ‰ •µc•jpµø äp쾫Āø•¾µì. 5 Iterate Aáá«Ğ «pa䵕µ‰ì ø¾ pµaµcp ³¾µ•ø¾ä•µ‰ aµj äpìá¾µìp.
8. Enhancing Runbooks Through AI Collaboration AI øäaµìˆ¾ä³ì ìøaø•c äĀµb¾¾¨ì •µø¾ jеa³•c, •µøp««•‰pµø äpì¾Āäcpì bĞ aĀø¾³aø•ca««Ğ ‰pµpäaø•µ‰ áä¾cpjĀäpì aµj bpìø áäacø•cpì ˆä¾³ •ìø¾ä•ca« •µc•jpµø jaøa. KpĞ bpµpˆ•øì: AĀø¾-Āájaøpì ˆä¾³ •µc•jpµø äp쾫Āø•¾µì Iµøpäacø•ėp øä¾Āb«p쐾¾ø•µ‰ ‰Ā•jaµcp Rpa«-ø•³p •³áä¾ėp³pµø ìĀ‰‰pìø•¾µì Iµøp‰äaøpj øpa³ ¨µ¾Ę«pj‰p T•ì aááä¾ac pµìĀäpì äĀµb¾¾¨ì ìøaĞ cĀääpµø aµj c¾µø•µĀ¾Āì«Ğ •³áä¾ėp øä¾Ā‰ AI-Ā³aµ c¾««ab¾äaø•¾µ.
9. AÈӌaÃpg Y“ȵ W“µƒf†“ß ßpÃm AI 1 Aa†åîe Ta»ƒ» Ijpµø•ˆĞ äpápø•ø•ėp Ę¾ä¨ˆ«¾Ęì ø¾ aĀø¾³aøp 2 Acce»» GPT BÈp†deµ LaĀµc CaøGPT aµj ìp«pcø 'Cäpaøp a GPT' 3 C“fpgȵe Ba»pc» Spø µa³p, áĀäá¾ìp, aµj c¾äp bpaė•¾ä 4 Tµap & CȻӌpîe U᫾aj ³aøpä•a«ì aµj jpˆ•µp äpìá¾µìp áaøøpäµì 5 Te»Ã IÃeµacÃp“» RĀµ ìa³á«p c¾µėpäìaø•¾µì aµj aj¥Āìø 6 De²†“å & Smaµe PĀb«•ì aµj ìpø accpìì áp䳕ì앾µì
10. The AI-First Mindset O†d Œpd»eÃ: "H“ß ca I »“†Þe Ãmp» ²µ“b†eŒ?" Haäj Ę¾ä¨, cä•ø•ca« ø•µ¨•µ‰. Neß Œpd»eÃ: "H“ß ca AI me†² Œe †““ƒ »ŒaµÃeµ ßmp†e »“†Þpg Ãmp» ²µ“b†eŒ?" Eˆˆ•c•pµcĞ, ¾áø•³•Ĩaø•¾µ, •³áäpìì. BÈp†dpg a c“††ab“µaÃpÞe fÈÃȵe W¾ä¨•µ‰ ø¾‰pøpä ˆ¾ä bpøøpä 쾫Āø•¾µì!
11. Prompting Like a Pro 1 Ask One Question Søaäø ʕø a 앵‰«p, ˆ¾cĀìpj ãĀpìø•¾µ ø¾ pìøab«•ì øp ˆ¾Āµjaø•¾µ ¾ˆ øp ìápc•ˆ•caø•¾µ. 2 Build on Previous Answers Eac ìĀbìpãĀpµø ãĀpìø•¾µ 쐾Ā«j «pėpäa‰p pa䫕pä äpìá¾µìpì ø¾ jpėp«¾á jppápä Āµjpäìøaµj•µ‰. 3 Iterate Toward Detail P侉äpìì•ėp«Ğ äpˆ•µp øp ìápc•ˆ•caø•¾µ øä¾Ā‰ øaä‰pøpj ˆ¾««¾Ę-Āá ãĀpìø•¾µì. 4 Finalize Developer Spec Sеøpì•Ĩp øp accĀ³Ā«aøpj •µˆ¾ä³aø•¾µ •µø¾ a c¾³áäppµì•ėp jpėp«¾á³pµø ìápc•ˆ•caø•¾µ.
12. Entering critical mode 1 Basic Review: "What potential edge cases am I missing?" Søaäø ʕø ãĀpìø•¾µì øaø ca««pµ‰p aììĀ³áø•¾µì 2 Deep Analysis: "How would this code behave under X conditions?" Eĝ᫾äp ìápc•ˆ•c ìcpµa䕾ì aµj •³á«•caø•¾µì 3 Architectural Review: "Could you help me think through alternative approaches?" F¾ìøpä c¾««ab¾äaø•ėp áä¾b«p³-쾫ė•µ‰ j•a«¾‰Āp
13. Embrace the Future, Engineers! 1 HȌa» gÈpde Ãme AI €“ȵeå 2 AI aŒ²†pfpe» mȌa cµeaÃpÞpÃå 3 T“geÃmeµ ße'µe ȍ»Ã“²²ab†e Tp ˆĀøĀäp •ìµ'ø ab¾Āø äpá«acp³pµø4•ø'ì ab¾Āø áaäøµpä쐕á. Aì pµ‰•µppäì, Ęp b䕵‰ øp ė•ì•¾µ aµj cäpaø•ė•øĞ, ʐ•«p AI b䕵‰ì øp c¾³áĀøaø•¾µa« á¾Ępä. Lpø'ì Ęp«c¾³p AI aì a c¾««ab¾äaø¾ä aµj ìpp ʐaø Ęp caµ ac•pėp ø¾‰pøpä!