Return-Path: Delivered-To: alvarochf@traceseis.com Received: from cs510.bluehost.com by cs510.bluehost.com with LMTP id 6HXPBAr682lylwAAWiBQxw (envelope-from ) for ; Thu, 30 Apr 2026 18:55:38 -0600 Return-path: Envelope-to: alvarochf@traceseis.com Delivery-date: Thu, 30 Apr 2026 18:55:38 -0600 Received: from wrqvxhpw.outbound-mail.sendgrid.net ([149.72.160.57]:53572) by cs510.bluehost.com with esmtps (TLS1.3) tls TLS_AES_128_GCM_SHA256 (Exim 4.98.2) (envelope-from ) id 1wIcAL-00000000Bu6-2kzm for alvarochf@traceseis.com; Thu, 30 Apr 2026 18:55:38 -0600 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=outlier.ai; h=content-transfer-encoding:content-type:date:from:mime-version:subject: to:list-unsubscribe:list-unsubscribe-post:cc:content-type:date:from: subject:to; s=s1; t=1777596925; bh=I32hDV6u6X5dwSW9Cw6/cwQAXfcJ41TzVJUXXvPINq0=; b=QusNwdl46mhCbVzecjGyjSFX9OABy71R6XrmmQzeXIUGfjEWLRTYFeY75P5m9hl/Cf3h x+YU7xXz6qoxJisLbiDr6eP2XiopFW5fVx4qglKhXAYpNrvrq72+/NGRiacznp+rVMqV7q fyS4Vx8/RdfmXFJQUenNFKYENjHeFd4IaZl0QuXMeLHsDvFdTQpqUwvo+PVoykFFOxb+HQ 3g0R5OhMa2XtpXa3BkU6L6cuCJKCNr0FVrCB5c2Iu7/FE87r25fHdnEzi+Tiy2ZQRWn+/P bajOfaWAjXFRnJzUe0zAnwe3SDyLS1VdclBKmK0hrmwrN8N6FZA3ZXUhzyUOf0Tg== Received: by recvd-697d59dcdd-cgm8w with SMTP id recvd-697d59dcdd-cgm8w-1-69F3F9FD-A1 2026-05-01 00:55:25.843551865 +0000 UTC m=+184693.582234346 Received: from MjIxMjY0MQ (unknown) by geopod-ismtpd-10 (SG) with HTTP id DUTZdfSXQ3anbgdIdYTV2Q Fri, 01 May 2026 00:55:25.834 +0000 (UTC) Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=utf-8 Date: Fri, 01 May 2026 00:55:25 +0000 (UTC) From: Outlier AI Mime-Version: 1.0 Message-ID: Subject: Inside Outlier: My dog taught me prompt engineering X-SG-EID: =?us-ascii?Q?u001=2EVMPt0=2Fm9kto67tqHKPmu5ER0QmDFniBt8B=2FzZMPq9ipJyiIGK7bg83bqP?= =?us-ascii?Q?=2FBNhcCnsCHuD09s9o7KX5Bq4jbGCRpJLr1BpfLn?= =?us-ascii?Q?xEv41gRVwBeZSMlkPiCiOqJg3ZXuu7SDAEEHcOZ?= =?us-ascii?Q?CuZfvdK57XtBzL68+8rP47I1+yJ03GlNvJAT74W?= =?us-ascii?Q?OTIzUG0DnA3zBzdqPDQSZRdPlqYSrUk9qcDRxlL?= =?us-ascii?Q?a=2FXhBMGj6GuPZuV3BZaU1ma1J8GVG=2F8Xt8adKGc?= =?us-ascii?Q?8OFC?= X-SG-ID: =?us-ascii?Q?u001=2ESdBcvi+Evd=2FbQef8eZF3Bn0=2F1wqYKdPmI+SecjXfzkcN8vxv0NICT+cfd?= =?us-ascii?Q?0lQ=2F3xd7pYAvHT7KRDKx3ATd7n+jh6OG406jlYA?= =?us-ascii?Q?eplfRltL4tpO5DpAfzXgDdvLXQCPpYGNkfkgEsg?= =?us-ascii?Q?4NQ1uViKejVCon+C6p3bLY41LG0uHfoIO1ik=3D?= To: alvarochf@traceseis.com X-Entity-ID: u001.iC7Op0uP6N8VlGsEj+vFyQ== List-Unsubscribe: =?us-ascii?Q?=3Chttps=3A=2F=2Fu2212641=2Ect=2Esendgrid=2Enet=2Flu=2Funsubscribe=3Foc=3Du001=2EYvI-2?= =?us-ascii?Q?u1H3DueMolqBrwRWvNcccnnGOKIFinWCo4Oi51E?= =?us-ascii?Q?UMvx5ArBmDZpjhI6vCqw3mNFbRMZC2CGpy8nd7o?= =?us-ascii?Q?cVXqDH6TUtbpMjjKIh60=5Fvu4j7NIwlDi-KIddh5?= =?us-ascii?Q?7j4nryw6Mb5ZDA9zULvQ9T9qnrItOi3JimAAVRR?= =?us-ascii?Q?L1xaLjChxio01M=5FUkI1y2ZMaBC7VeUFrpmwy9AL?= =?us-ascii?Q?ypZrkpTE-1kKdKIbJfRY3gDvXrWEwLC3Z7o4GI3?= =?us-ascii?Q?55UdzSorWf=5F0P6RNTc6=5F6viRqrKs6VBPAY19xGO?= =?us-ascii?Q?Cr=5FU15fMunnU3pFzVJPBNa6jkpDf8EfvHubJWmO?= =?us-ascii?Q?UpS6b8jXicA1x4X1nFaCGRE2Uqo0dP2eWYJhnkl?= =?us-ascii?Q?vkZlhwLV7C6KHT8vJCpF7f4XHzePmRF8gmsubmE?= =?us-ascii?Q?9GlA26HzxD5EPAbpAXeTnbAj8NHJsY1jmozK4F2?= =?us-ascii?Q?5-DClXekIS=5FP-k6e4y5iYOUoJsQR9-wpDp1GslC?= =?us-ascii?Q?dyval2-=5FYI6clqBvc5gAyPkjIk3HSrGIYz3cBNS?= =?us-ascii?Q?Vn2WaOfsg3S1XP8zFMneevfKf3mJwdoJfuQ9afn?= =?us-ascii?Q?nJoEmrI9FmObJa=5Fv-l9f0aYoly4oi2rU91Q-FIj?= =?us-ascii?Q?U3SrxSLAvUzHuoAkZ54HsoOUxTR4hU5SlTbiB8o?= =?us-ascii?Q?1-FgxBD0ca6cvng7MdIJvHRYFiJECmSpwzDxqsF?= =?us-ascii?Q?wfM7SeWespguN4zXv5ZGRL71uVc2EFNVbb08vwk?= =?us-ascii?Q?5T6d6NDpjvNGPOTgURtMlBySl-6h0-fJuI=5FJgsq?= =?us-ascii?Q?j6GxaA7UlDSRb7lPB1augtfnsKMWyfhTrx0p9qa?= =?us-ascii?Q?aMPDYLDgshZaZ6utWGAuzpxn2Q=3E=2C=3Cmailto=3Auns?= =?us-ascii?Q?ubscribe=40em528=2Eoutlier=2Eai=3Fsubject=3D22126?= =?us-ascii?Q?41=2EkBOKDmC5-d4kiUvIe0SjGL=5F-jIb-RroEG0G=5F?= =?us-ascii?Q?FuBs9m1oMDAwdTAwMFdDyUziU9HSZ54JUp0bzTf?= =?us-ascii?Q?UyF-x6M8ABdnITs8LdtO1oyxqcsMTiIDhSDo1Nx?= =?us-ascii?Q?rxWUiiJDv3sT-5DcSf0Q8wD=5FhSH-96WcKTJXf7R?= =?us-ascii?Q?Df84Qqck9g8oxdpi0FUnVOEIpmuMw=3D=3D=2ECXM8CZw?= =?us-ascii?Q?LVROo6S4RDmBzzs69pOwegqAPMjYK=5FMQ64UNoMD?= =?us-ascii?Q?AwdTAwMA=5FrVESp-IQti5n17h-a58-7jpKI7kKiC?= =?us-ascii?Q?l=5F19EW-U9NcvYExA8IX9WaBD=5FnEqKj53cJZKGqL?= =?us-ascii?Q?J4UsoSWiwcoM1DR4qi9BWsZQjpLNzVdhvuQcHgY?= =?us-ascii?Q?KXDjJrJCoUxNHL=5Fe87VfcTiKyEMCcf0RtFiUd7V?= =?us-ascii?Q?hjM6GgQiDPB3sU7DVC7KdqAz5izuQwnPJ89USue?= =?us-ascii?Q?mMzWOj=5Fi7lt0EKghuSBcmOkMV9n5CGoqwyGg4wR?= =?us-ascii?Q?YqK76GlMEwOBJpHfXNmO-6=5FFHMhK30pLtaxIDne?= =?us-ascii?Q?y2ihMehRRCzUGzA4TvDIe6CRh5n9u-dws=3E?= List-Unsubscribe-Post: List-Unsubscribe=One-Click X-Spam-Status: No, score=0.3 X-Spam-Score: 3 X-Spam-Bar: / X-Spam-Flag: NO

=20 =20

<= /p>

TL;DR:

  • Six prompt engineering = tips that will change your outputs =E2=80=94 Take the free course
  • How to write support t= ickets that get resolved without the back-and-forth
  • How has your AI communication change= d in the past year? Let=E2=80=99s discuss.

    • Hey, it's Eliana.

    Squiggles does t= his thing where he ignores you unless you're extremely specific.

    =

    "Com= e here" =E2=80=94 nothing.

    "Come here please" =E2=80=94 still nothing= .

    "Come here please so I can give you a treat" =E2=80=94 immediate co= mpliance.

    It's not because he suddenly understands me better, it's be= cause he's conditioned. Certain words map to certain outcomes, and he's lea= rned which ones matter.

    That's basically what's happening when you pr= ompt a model.

    It's not reading your mind. It's not filling in your in= tent. It's responding to patterns it's seen before, and your job is to hit = the pattern that gets you what you want.

    Vague prompt =E2=86=92 gener= ic output.

    Specific prompt =E2=86=92 higher chance you trigger the ri= ght pattern.

    Same underlying system. Different species.

    = Prompt Engineering Tips & Tricks

    Whethe= r you're prompting an AI or training your dog, the principles are the same:= clear inputs lead to better outputs. Here are some of the best tips we've = picked up:

    1. Lead with the outcome, not the task. Don't just say what you want done =E2=80=94 say what a= great result looks like. "Write a summary" is weak. "Write a 3-sentence su= mmary a 10-year-old could understand" is strong.
    2. Give it a role. Starting your prompt with something li= ke "You are an expert in X" helps set the tone and context. It focuses the = response.
    3. U= se examples. Show, = don't just tell. If you want output in a certain style or format, give an e= xample of what that looks like. One good example is worth ten sentences of = explanation.
    4. Want bullet points? A numbered list? A table? Say so.
      5. <= li>Iterate, don't restart.= If you don't love = the first response, don't scrap everything. Add a follow-up: "Make it short= er," "Make the tone more casual," or "Focus only on X." Prompting is a conv= ersation.
    5. E= liminate ambiguity.= Read your prompt back to yourself. If it could mean two different things, = the AI will probably produce the wrong one. Rewrite until there's only one = possible interpretation.

    Want to go deeper? Our Prompt Engineer= ing Course breaks all of this down step by step =E2=80=94 with real example= s, exercises, and frameworks you can use immediately.

    =F0=9F=91=89 = Check out the free course here

    Writing Support tickets that get = resolved fast

    When something goes wrong on Outlier, a clear ticket i= s the fastest path to a fix. The more we can act on without asking follow-u= p questions, the sooner your issue gets resolved.

    Start with the bas= ics: your account email, task ID or project name, and what happened. Missin= g basic info almost always triggers a follow-up from us, which can add a fu= ll day to your wait before we can even start looking into the issue.=

    = Then be specific. "My submission wasn't approved" gives us almost nothing t= o work with. "I submitted task #12345 on April 28 and it's been in review f= or 48 hours" gives us everything we need.

    If something looked wrong o= n screen, attach a screenshot rather than describing it. If you already tri= ed refreshing or clearing your cache, mention it so we don't retrace your s= teps.

    If you have two separate problems, open two separate tickets. I= t keeps things from getting tangled and makes sure nothing gets missed.

    What you've been = learning

    Nomvelo is working on communicating under pressure =E2= =80=94 in her words: "Saying what I mean clearly, staying calm, and not ove= rthinking every word after I've said it. It sounds simple, but it's surpris= ingly hard." (It is. It might also be the most transferable skill anyone ca= n build.)

    Raiyan Ali is learning programming = and problem-solving =E2=80=94 building on the AI analysis skills he develop= ed at Outlier.

    =EC=A1=B0=EC=9C=A4=EC=A0=95 is studying for her professio= nal organizer certification. She loves the feeling of a clean, organized sp= ace, and she wants to give that to other people. Plans to freelance once sh= e passes. Her last line: "I want to do good for myself and have a positive = influence on my neighbors." I just love that!

    Hit reply

    <= p>How= do you talk to AI differently than you did a year ago?

    My answer: = a year ago I described what I wanted. Now I explain why I want it, what I'm= going to do with it, and what I've already tried. The first response is al= most never the best one. Reply here, or post your answer in = Outlier Community. 

    Time to walk my favorite prompt engineer,

    Eliana

    =20
    Our privacy policy can be found at https://outlier.ai/l= egal/privacy-policy. If you have questions about this Privacy Policy, pl= ease contact us at privacy@outlier.ai or by mail at: Attn: Privacy, 650= Townsend, San Francisco, CA 94103.
    =20
    3D""