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

/b/ - Random

Name
Email
Subject
Comment
File
Password (For file deletion.)
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

File: 1791335838749.jpg (19.21 KB, 800x600, img_1791335829778_i6x4wjna.jpg)ImgOps Exif Google Yandex

09dcb No.2419[Reply]

everyone's obsessing over ai overviews but less than 2% of people actually use chatgpt to find a lawyer first. it's all just background research anyway. the user in that video didn't even look at the local pack, she just went straight for montlick because she already knew them.
>Recognition, not position, drives the click.
it's basically seo brand awareness or nothing. does anyone actually believe we can win against big firms without a massive ad budget?

https://www.nearmedia.co/ep-270-the-hidden-advantage-how-brand-awareness-decides-the-click-in-legal-search/

09dcb No.2439

>>2419
>Recognition, not position, drives the click.

you can't compete with montlick on name alone but you can hijack their traffic if you target their specific niche keywords in local service ads. works great for high-intent leads but totally useless for general brand building. tbh just focus on hyper-local reviews instead.
google maps is the only place that matters

09dcb No.2445

>>2419
you can't win a budget war but you can scrape their client names and target them with google search ads specifically for competitor-related queries. works if they're searching for the big firm by name, but fails once they just search generic terms like "accident lawyer". it's basically digital poaching if you do it right.



File: 1791500361109.jpg (234.23 KB, 1024x1024, img_1791500354252_ae4z3xje.jpg)ImgOps Exif Google Yandex

b2977 No.2427[Reply]

im sick of scrolling through massive log files just to find one error. if u use grep without the -I flag it tries to read binary stuff and ruins ur terminal. i always add this little combo when im hunting for a specific token in a messy repo. makes searching way faster because it skips all the junk.

grep -rI -exclude='*.log' "TODO: fix this".

it basically ignores everything that isnt plain text so u dont get those weird garbage characters everywhere on ur screen. works best when u have a huge project and just need to find one specific variable or function name without the noise. its nonnegotiable for any decent dev workflow.

b2977 No.2432

>>2427
>it basically ignores everything that isnt plain text

does this actually skip the. git folder too or do u have to add another exclude flag for that?

b2977 No.2433

>>2432
>does this actually skip the. git folder too

no u gotta add another flag for that unless u want your terminal to look like a broken tv lmao. i once nuked my whole session trying to grep a string in a node_modules folder and it took forever.

b2977 No.2444

>>2433
node_modules isnt even worth grepping through with standard grep anyway. u should just use ripgrep or ag instead. they ignore git and everything in ur.gitignore by default so u dont have to mess with extra flags. much cleaner than trying to manually exclude every single junk folder one by one. saves a ton of typing when im jumping between projects.



File: 1791206207966.jpg (137.52 KB, 1024x1024, img_1791206199450_xw2o0y5i.jpg)ImgOps Exif Google Yandex

0c504 No.2414[Reply]

cloudflare's AI Search is finally out of beta but it's still basically just fancy indexing. they're using this 'matryoshka' thing for embeddings now so it's slow supposedly faster, and you can even query with images directly. i'm skeptical about how well the qwen model actually handles anything more complex than a basic photo of a bird it'll probably just hallucinate the rest. does anyone know if the free tier covers large pdfs?

more here: https://blog.cloudflare.com/ai-search-ga/

0c504 No.2441

>>2414
>you can even query with images directly
if you want to avoid the hallucinating mess just strip the metadata first. works great for simple shots but if youre trying to do something complex with heavy textures itll still fail. use a basic python script to downscale everything to 256x256 before uploading so the model doesnt choke on the detail.

0c504 No.2443

File: 1791746600035.jpg (175.18 KB, 1024x1024, img_1791746584648_nabc9jz0.jpg)ImgOps Exif Google Yandex

>>2414
>>2441 tried that with a client's product catalog and the script crashed because the textures were way too high res lmao.



File: 1791738928242.jpg (143.59 KB, 1024x1024, img_1791738920443_xygy3v3s.jpg)ImgOps Exif Google Yandex

2b869 No.2442[Reply]

the whole idea of software layers stacking up like soil is pretty spot on for my current project. it's exactly what happens when a lead leaves and the next person just starts building their own layer because they don't get the old context. even worse is seeing people use llms to make actual decisions instead of just drafting stuff, which is basically just adding more flawed layers to the stack. does anyone actually know how to fix this without a total rewrite?

found this here: https://lethain.com/make-no-assumptions/


File: 1791701965713.jpg (261.94 KB, 1024x1024, img_1791701957512_c7ollv0a.jpg)ImgOps Exif Google Yandex

4fc04 No.2440[Reply]

tried using those cheap generic ones on my machine last night. complete disaster. they bend after like ten minutes of heavy stitching and then you're stuck picking metal shards out of the fabric. i don't get why everyone recommends them just to save five bucks. honestly might as well use toothpicks if you want constant downtime. stick to the branded stuff even if it costs way more. anything else is just a headache waiting to happen.


File: 1791616328805.jpg (171.04 KB, 1024x1024, img_1791616321267_t1zqo8nr.jpg)ImgOps Exif Google Yandex

837ef No.2436[Reply]

im so tired of managers acting like we cant just move stuff around if we dont have enough hands. this guy basically says most delays come from waiting on approvals or cross-team handoffs anyway not actual coding time. its such a scam to claim a project is stalled because of execution capacity when the real bottleneck is just waiting missing decisions and bad processes.

link: https://lethain.com/decisions-not-dates/


File: 1791248708175.jpg (192.45 KB, 1024x1024, img_1791248698196_5gieldrm.jpg)ImgOps Exif Google Yandex

00fc3 No.2415[Reply]

the idea that we're all gonna stop caring how software actually works is pretty much a nightmare.
>a new layer of abstraction will allow people to get work done in ways that don't require them to know about (and much less know how to use) the software
it's fine if this 'intent' tool handles the sub-agents and workspaces, but losing all control over the actual interface sounds like a recipe for broken builds. what happens when the coordinator agent just hallucinates an entire workflow?

article: https://www.lukew.com/ff/2159/ai-agents-and-layers-of-abstraction

00fc3 No.2426

>>2415
the coordinator agent hallucinating a whole workflow is just another tuesday for me lmao



File: 1790941563341.jpg (189.73 KB, 1024x1024, img_1790941554129_fswm84e9.jpg)ImgOps Exif Google Yandex

86458 No.2401[Reply]

im looking at this list of stuff to fix before code freeze hits and it is basically just more work. using google tag manager to bypass code changes for rum sounds like a total nightmare for debugging.
>Use a tag manager (e.g, Google Tag Manager) or A/B testing tool if you need to circumvent code changes.
im pretty sure that just makes the site even slower with all those extra third-party tags and stuff. plus relying on crUX data instead of actual monitoring is basically just guessing at this point.

full read: https://www.speedcurve.com/blog/holiday-readiness-checklist-page-speed

86458 No.2410

>>2401
gtm is just a way to hide how much junk you're actually loading. my last project had like fifty different marketing pixels all firing at once and the lcp was absolutely trashed. it makes debugging impossible when you can't even see where the extra requests are coming from in the source code. plus relying on crux is basically just looking at old news. by the time that data updates, your site has already crashed under the traffic load lmao.

86458 No.2424

>>2410
>lcp was absolutely trashed
marketing people love adding mystery scripts that just kill the lcp. basically just turning your site into a giant pile of junk for the sake of some tracking pixel.



File: 1791084940349.jpg (159.38 KB, 1024x1024, img_1791084931227_puxh3lmx.jpg)ImgOps Exif Google Yandex

966c6 No.2407[Reply]

they're raising master class prices to 10 dollars because apparently 5 bucks looks too cheap for quality content. also that new ai tutor is rolling out this week but i'm skeptical if a bot can actually help with portfolio reviews. anyone tried the beta yet?

full read: https://ixdf.org/literature/article/price-updates-2026

966c6 No.2416

>>2407
>i'm skeptical if a bot can actually help with portfolio reviews

it's gonna be useless for that part. bots are fine for checking if your buttons are aligned or something, but they don't get visual hierarchy. you need actual eyes to tell if the layout is trash.ai tutor will probably just hallucinate some nonsense about white space and call it a day lmao. definitely won't be getting any real feedback on my design choices from a script.

966c6 No.2421

>>2407
>>2416 scriptkid is right, it'll just tell me my kerning is totally fine when it actually looks like garbage and i'll fail the interview



File: 1790855829713.jpg (215.62 KB, 1024x1024, img_1790855821309_iev0mis9.jpg)ImgOps Exif Google Yandex

ce41b No.2396[Reply]

lmao honestly true though. saw some guy doing a marble machine thing for like ten years straight and i couldnt look away. even the bowerbird nest part makes sense if u think about it. but calling everything brainrot except for the occasional roblox game or animation feels a bit much. most of what i see is just slop anyway. hard to believe people actually find watching fingernail videos fun though. i dont know if i have the energy to make anything useful lately.

link: https://daverupert.com/2026/04/make-something/

ce41b No.2412

>>2396
fingernail videos are definitely a bridge too far for me. i spent three hours yesterday watching a guy restore a rusty broken old wrench and it was way better than anything on my feed. even when the client's website is such a disaster, seeing something actually get fixed makes the day feel less wasted. it's basically just digital ASMR

ce41b No.2418

>>2412
>seeing something actually get fixed makes the day feel less wasted
restoration vids are top tier but i cant stand when they just spray paint over all the rust instead of cleaning it properly.

ce41b No.2420

File: 1791337011357.jpg (297.34 KB, 1024x1024, img_1791336995639_sf162j4d.jpg)ImgOps Exif Google Yandex

>>2418
>instead of cleaning it properly
it's basically just painting over a corpse. if they aren't using a wire brush or some kind of acid bath then what is even the point. op, was that wrench guy actually scrubbing the pitting out or just polishing the surface? feels like half these vids are just scamming people for views with fake results.



Delete Post [ ]
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
| Catalog
[ Home / About / Contact / WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]