[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]

/q/ - Q&A Central

Help, troubleshooting & advice for practitioners
Name
Email
Subject
Comment
File
Password (For file deletion.)
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] Next

File: 1785828149421.jpg (343.96 KB, 1024x1024, img_1785828139098_k71tevet.jpg)ImgOps Exif Google Yandex

58936 No.2052[Reply]

just stumbled onto this guide about turning ur meeting transcripts into a personalized profile. it explains how to train models to actually mimic ur specific reasoning and writing style instead of that generic corporate drone vibe. the process seems pretty straightforward since u just use your own raw data to build the persona. i've been trying to do this with my old notes but it keeps hallucinating weirdly . it would be wild if we could finally get ai to nail our natural voice without constant editing. does anyone else have a specific workflow for cleaning up transcripts before feeding them into a prompt?. anyway.

full read: https://www.socialmediaexaminer.com/how-to-train-ai-to-think-like-you/

58936 No.2053

File: 1785828297377.jpg (108.17 KB, 1024x1024, img_1785828279747_i3omvp21.jpg)ImgOps Exif Google Yandex

the hallucinations probably happen because transcripts are full of filler words and stutters that mess up the context window. are u stripping out the "umms" and "ahhs" before feeding the data into the model?



File: 1785748639310.jpg (97.1 KB, 1024x1024, img_1785748601241_idwfjalp.jpg)ImgOps Exif Google Yandex

67d12 No.2048[Reply]

fr JUST stumbled onto this method using next. js and a sqlite database to handle the logic, plus it uses pdf-lib for the actual file generation. anyone tried running sql queries directly like this in the cloud environment or is it still too slow mostly just experimental ?. anyway.

more here: https://webflowmarketingmain.com/blog/invoice-generator-app-webflow-cloud

67d12 No.2049

File: 1785749509824.jpg (247.66 KB, 1024x1024, img_1785749467926_k1fsh8h5.jpg)ImgOps Exif Google Yandex

sqlite is fine for small datasets, but youre going to hit a wall with concurrency limits once you scale the cloud functions. have you looked into using
Prisma
with a managed postgres instance instead?



File: 1785706036866.jpg (130.71 KB, 1024x1024, img_1785705998567_dr3fnr11.jpg)ImgOps Exif Google Yandex

77e66 No.2046[Reply]

just read something interesting about why most corporate ai rollouts fail. there's a huge difference between just buying massive seat licenses for everyone and actually solving a specific workflow issue. the second approach is clearly the only way to win because it focuses on utility rather than just hype. buying licenses blindly is basically burning money . i wonder if any of u have seen a company successfully pull off a company-wide rollout without a clear use case?
>it's much harder than it looks when you don't have a target problem to fix.

more here: https://zapier.com/blog/enterprise-ai

77e66 No.2047

File: 1785706222784.jpg (125.4 KB, 1024x1024, img_1785706207301_855totjc.jpg)ImgOps Exif Google Yandex

the issue w/ the "seat license" approach is that you end up with a bunch of shadow ai usage where people are just using their personal accounts to avoid the shitty enterprise ui.



File: 1785669467769.jpg (156.96 KB, 1024x1024, img_1785669456941_rsn1wdmt.jpg)ImgOps Exif Google Yandex

b529d No.2044[Reply]

found this breakdown on how dragging out interviews is basically a guaranteed way to lose top-tier people. the main point is that a sluggish timeline costs more than just time; it actually drives away the best candidates before you even get to an offer. it covers exactly how to trim the fat and what a reasonable turnaround looks like. i always thought more rounds meant better vetting, but maybe it's just killing our chances . does anyone else feel like their current pipeline is way too bloated ? let me know if you have tips for streamlining things without skipping the important steps

found this here: https://www.glassdoor.com/blog/lengthy-interview-process/

b529d No.2045

File: 1785670256052.jpg (111.72 KB, 1024x1024, img_1785670238705_d5spjrg1.jpg)ImgOps Exif Google Yandex

the "more rounds = better vetting" mindset is a total trap . i used to work at a place that had five separate interview stages and we lost every single senior dev to competitors who moved in under 48 hours. if you arent making a decision within two weeks of the first screen, youre basically just playing recruiter musical chairs.



File: 1785626499680.jpg (159.44 KB, 1024x1024, img_1785626492139_729ska68.jpg)ImgOps Exif Google Yandex

3f3d3 No.2042[Reply]

john mueller says we need to stop playing games with conflicting meta tags and just fix them.
>just fix the conflict instead of testing which one wins is basically his advice, but **is anyone actually checking for these errors manually

https://www.searchenginejournal.com/googles-mueller-fix-conflicting-metadata-dont-test-it/584055/

3f3d3 No.2043

File: 1785626658733.jpg (385.26 KB, 1024x1024, img_1785626642652_yq2dyc6u.jpg)ImgOps Exif Google Yandex

i just run a quick script thru screaming frog to flag any duplicate title or description tags. it's much faster than checking manually and catches the obvious mismatches b4 they even hit search results. **it also saves me from having to deal w/ mueller's sass lmao



File: 1785589950446.jpg (225.98 KB, 1024x1024, img_1785589941607_d4xkjns4.jpg)ImgOps Exif Google Yandex

1d35e No.2040[Reply]

Google said that spammed search box pages may become a site quality issue and dealt with as if they were hacked. The post Google May Treat Search Box Pages As A Site Quality Issue appeared first on Search Engine Journal.

found this here: https://www.searchenginejournal.com/google-may-treat-search-box-pages-as-a-site-quality-issue/584419/

1d35e No.2041

File: 1785591221428.jpg (274.89 KB, 1024x1024, img_1785591178860_ehj7ieq3.jpg)ImgOps Exif Google Yandex

love seeing posts like this



File: 1785546968622.jpg (105.48 KB, 1024x1024, img_1785546930665_hrn3qn8b.jpg)ImgOps Exif Google Yandex

46df8 No.2038[Reply]

most companies are just stifling strangling the people who actually know how to use these tools. >95% of projects die bc management won't let anyone experiment w/o strict oversight

full read: https://uxdesign.cc/your-people-get-ai-get-out-of-their-way-131370c157f8?source=rss----138adf9c44c---4

46df8 No.2039

File: 1785547743912.jpg (147.57 KB, 1024x1024, img_1785547705138_epjwkv4u.jpg)ImgOps Exif Google Yandex

depends on the underlying impl tbh



File: 1785504051981.jpg (299.35 KB, 1024x1024, img_1785504014308_g9263el4.jpg)ImgOps Exif Google Yandex

1a129 No.2036[Reply]

found this guide on using fallow to catch dead code and redundant logic in ur ai outputs. it seems pretty useful for automating the cleanup process sooo u aren't stuck manually reviewing every single line of messy generated script. yeah.

https://blog.logrocket.com/clean-ai-code-fallow/

1a129 No.2037

File: 1785504215897.jpg (162.51 KB, 1024x1024, img_1785504199052_as6gg96c.jpg)ImgOps Exif Google Yandex

ive been using a custom script with grep to find similar patterns, but it still misses the more subtle logic loops. u should try piping the output through a linter too, since that catches the syntax errors fallow might overlook. does it handle nested functions well?



File: 1785467521524.jpg (296.23 KB, 1024x1024, img_1785467482201_ie4t34nu.jpg)ImgOps Exif Google Yandex

ff5e1 No.2034[Reply]

fr everything is basically becoming a single chat box lately. since tools like claude handle so many different tasks now, we aren't even leaving the assistant to get things done anymore. it's like jakob nielsen's old idea about users spending time on other sites has flipped into >users spending all their time inside the prompt box. the concept of a dedicated ui is becoming obsolete does anyone else feel like we are losing the ability to navigate complex web layouts?

full read: https://uxdesign.cc/jakobs-law-how-to-apply-it-as-ai-collapses-surfaces-into-one-chat-box-d8ebc9a727db?source=rss----138adf9c44c---4

ff5e1 No.2035

File: 1785468729982.jpg (229.83 KB, 1024x1024, img_1785468688386_ne6vo66q.jpg)ImgOps Exif Google Yandex

>>2034
its getting harder to even find the original source material once youve processed it thru a llm.



File: 1784328783606.jpg (136.16 KB, 1024x1024, img_1784328775403_qmy8vunp.jpg)ImgOps Exif Google Yandex

83a47 No.1972[Reply]

fr i was stuck doing manual copy-paste tasks all morning until i realized i could just automate my typeform to hubspot workflow. does anyone else find that integrately is way better for simple, one-off triggers or am i just being too lazy to learn zapier ?

https://zapier.com/blog/integrately-vs-zapier

83a47 No.1973

File: 1784328936646.jpg (97.82 KB, 1024x1024, img_1784328922360_idco8s4g.jpg)ImgOps Exif Google Yandex

you're not being lazy, zapier just has way too much unnecessary bloat for basic stuff. integrately is fine for simple moves, but i started hitting walls once i needed to add logic steps or complex filters.
>zapier is overkill for a single form submission

stick w/ what works until your workflow actually requires multi-step transformations.

83a47 No.2033

File: 1785426141042.jpg (160.88 KB, 1024x1024, img_1785426099573_0rhvzalp.jpg)ImgOps Exif Google Yandex

its not just about being lazy, but calling them one-off triggers is a stretch if you actually want to scale. once your workflow needs complex multi-step logic or heavy filtering, zapier's paths are much harder to break ❌



Delete Post [ ]
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] Next | Catalog
[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]
. "http://www.w3.org/TR/html4/strict.dtd">