https://posts.cv/kelsy/nMN8CX6U9p7raPm46AT5

https://chat.openai.com/share/52f389a1-792b-4f1b-b103-e22d71dd8c04

https://www.logitech.com/en-us/software/logi-options-plus.html

https://www.figma.com/community/plugin/1182746451598399442/atlas-icons

https://posts.cv/kelsy/AP0nhBEoRyBBa4NQsG7R

https://stability.ai/news/stable-video-diffusion-open-ai-video-model

microsoft.com/en-us/edge/download?form=MA13FJ

https://help.invisionapp.com/hc/en-us/community/posts/11525657213965-SUNSET-NOTIFICATION-Studio

https://are.na/kelsy-gagnebin/listening-s81h3jnxl68

https://gist.github.com/adispezio/063db7d25d659e177d152d1eac037e8d

https://blog.google/products/gemini/bard-gemini-advanced-app/

kickstarter.com/projects/flarup/the-macos-app-icon-book

https://amzn.github.io/style-dictionary/#/README

https://www.notion.so/product/calendar/download

https://youtu.be/3HhBpfyKH_M?si=TGcEKlMeVm9A3Wc5&t=1785

https://rosenfeldmedia.com/books/design-that-scales/

peek at last week && my apparent practice of adding images, only to remove them.

fun onboarding + solid results with just a single-word prompt (i forgot to finish typing)

the ability to remix different styles is rad && generating 4 images feels appropriate to me.

🤖🎨 https://visualelectric.com/

fun to poke around
🕹️ https://webamp.org/

&& the skins archive is definitely worth a look
🎨 https://skins.webamp.org/

arc max’s tidy downloads has been nice, except for dall•e images.

(prompt details are in the filename)

copilot (web) updated the example prompt format && the notebook tab is interesting.

it has a 18k char-limit (instead of 4k or 2k w/copilot) && doesn’t seem to keep history (also errs for me when trying to generate images)

currently working through a lot of bad ideas, in order to get to some better ones.

posting something i donotlike has actually increased my likelihood of fixing it.
(vs. hiding it forever on localhost)

poking around a new (to me) app for running LLMs locally.
💾 https://lmstudio.ai/

• ability to discover && download models
• lots of info to peek at
• links to model urls && arxiv papers

(if you prefer terminal, i’d peek ollama)
📄 https://posts.cv/kelsy/nMN8CX6U9p7raPm46AT5

running lighthouse after some cleanup + optimizing.
i was pretty delighted with the confetti (my first time seeing those numbers)
now i just need to make the actual content* less-shit-more-good (but it’s a continuous process)

(time-lapse of some of that cleanup in thread)

this doodle didn't start with a theme, but then i put some 🛼 on one && it seemed to make sense.

(last couple hours / more cleanup + getting tags to filter)
there's a couple chatgptQuestions towards the end.
i've noticed better results saying code will be between three backticks then ```paste_code```

🤖💬 https://chat.openai.com/share/52f389a1-792b-4f1b-b103-e22d71dd8c04

if you’ve got images you want converted into SVGs, this tool is free while in beta.
🖼️ https://vectorizer.ai/

(also contains a lot of great info on what the different settings actually do)

(peek at exploring ideas earlier today)

not saying any of my patterns are worth repeating, but it’s nice to be able to explore an idea across a range of tools && contexts w/out much thought.
(+ it's fun to explore w/ui-hidden)

less-thought took years of semi-thoughtful-practice

(peek back at monday)
nothing too exciting, just starting to slog through the gnarlyMessIveMade. (current site's code+content+design is somewhere between dogshit—dumpsterfire)
obsidian for high-level notes, notion for a scratchpad && basic spellcheck.

(this was an incredibly fast subscribe for me)
i'm only ~30mins in, but am really digging how detailed mark is getting, && am sincerely looking forward to the 7+ hours of figma_ui details 😍

📼 https://www.youtube.com/@designloomco

(sometimes things lineup)
spotted first backyard bee of 2024, received arc anniversary email, && my morning routine of doing nothing but feeling air go in&out my nose is a repeating digit.

if utah had the lottery, i’d be buying a ticket.

🎅-magic

the app has fun patterns for ai-image generation && the community is worth checking out.
🤖 https://app.recraft.ai/community

(just moving the hand around)

started with some &_practice, then some doodles to fill up the page.

great icon organizer got update that’s worth a peek
💾 https://nucleoapp.com/application

(rad free app, a two-person team, can pay for the icons && i’m a sucker for payBeforePriceIncreases + the upgrade from v01 was 50% off - which feels generous af)

the default ai-smart actions for summarizing with gpt4 saves some clicks.
(could tweak the prompt to get more refined responses)

💾 https://www.logitech.com/en-us/software/logi-options-plus.html

galileo is out of beta && did a quick figma test (it err'd for me in beta)
copied nicely + auto layout means moving things around is easy.

gallery cards showing prompts + personWhoPrompted && ability to remix/justCopyToFigma is pretty rad.

🤖 https://www.usegalileo.ai/

i’m starting to think that figma wants me to interact with its ai-features.

(i searched for cat in atlasicons && got yoga poses, which was fun)
💾 https://www.figma.com/community/plugin/1182746451598399442/atlas-icons

i like including url’s to things (even if it looks gnarly)

meta-example:
• copied frame w/live-text of url as png (cmd+shift+c)
• grabbed url-text from png w/app
• someone could grab url from this video

(transmitting info into unknown contexts 🏄)
💾 https://textsniper.app/

quick civic duty reminder to vote

(i'll be coming out with a masterclass on the figma pencil tool soon)

🗳️ https://posts.cv/icons

i always thought this was a cool project (there's more than just 90's)

📺 https://www.my90stv.com

pretty fun interactions
(not 100% sure if it’s worth signing up for, i feel like i just got on a shitload of email lists 🤷)

can download a replay of your session (mp4) 👍
img composition has watermark, but the single img doesn’t (png - 512x512) 👍

🤖 www.freepik.com/pikaso

nice open-source project for running LLMs locally (fun site + rad app && you can download models from it)
💾 https://jan.ai/

(like a nicer lm studio)
📄 https://posts.cv/kelsy/AP0nhBEoRyBBa4NQsG7R

poking around the latest-greatest with nuekit’s v0.1 release.

🎉 https://nuejs.org/blog/nuekit-010/

lots of reading docs, throwing on borders to see what i’m working with && just starting to figure out how things work.
(the past 2 hours condensed at 60fps)

excited for the future of generative video (the selectArea + type-to-change is very interesting)

🔗 https://pika.art/

(stable video diffusion was also released a couple weeks ago)
📄 https://stability.ai/news/stable-video-diffusion-open-ai-video-model

(just arrived 🎉, haven’t started)

i really like ethan mollick’s work - ‘research-based view on the implications of AI’
&& would recommend taking a peek at
substack:
📄 https://www.oneusefulthing.org

some prompts:
📋 https://www.moreusefulthings.com/prompts

i’d check out edge(mac) if you haven’t in a while
💾 microsoft.com/en-us/edge/download?form=MA13FJ
&& immediately update the start page 
🔗 https://copilot.microsoft.com/

built in copilot is pretty interesting

(link to ai_music_test from vid)
📻 https://sl.bing.net/iLZJbJCd7am

some things i liked about studio
• release notes show team
• motion editor + ability to pin side || bottom
• preview in new window

how not* to sunset enterprise (view comments)
📄 https://help.invisionapp.com/hc/en-us/community/posts/11525657213965-SUNSET-NOTIFICATION-Studio

making a more conscious effort to upload things i listen to + a chance to use some generative art as covers.
📻 https://are.na/kelsy-gagnebin/listening-s81h3jnxl68

(currently going back through alan watts' lectures, which might not be for everybody)

🙈 peek at doing some boring stuff.
some figma cleanup + upgrading to astro 4 (followed by build errs && panic-googling)

note-to-self:
double-check astro config if adding-then-removing a dependency.

📄 https://astro.build/blog/astro-4/

my nervous tick of hitting ⌘ + s every few minutes has finally dropped.

here’s the gist to copy-paste && peek your shortcut usage:
🔗 https://gist.github.com/adispezio/063db7d25d659e177d152d1eac037e8d

bard is officially gemini && there’s an option to try 'advanced'/ultra w/google one ai premium subscription.
(includes using gemini in gmail,docs,slides,sheets - formerly duet ai)

🔗 https://blog.google/products/gemini/bard-gemini-advanced-app/

remembering invision's sketch plugin, that had a badass web browser feature.

click on text && it would copy automagically to the selected text-layer (shapes would fill with selected images)

first quick tests w/pika (invite came today)

• used img generated w/google + my default test prompt, pretty fun result
• it got weird at the end when i lowered ‘consistency with text’ && increased 'strength of motion'
• (results w/out a starting img will require re-prompting)

last week was spent reinstalling OSX a few times (disk util. had an err it couldn't fix && i had trouble accepting my reformatReality - please backup your files)

but today a beautiful kickstarter project arrived 🎉
🌐 kickstarter.com/projects/flarup/the-macos-app-icon-book

google image generation test (same prompt as dalle•3)

the most interesting thing was the option to ‘help me write with more detail’

dalle seemed to expand from the original prompt, but this detail was about the generated image*

seems easy to go back && explore options

first test of my dall•e 3 access.

(impressed by the improved prompts && the initial results)

peek throwback (~2019) working through design tokens, but from the code-side first (tools/life was different)

style-dictionary would export to all platforms (we had web + android), which was really interesting at the time.

🌐 https://amzn.github.io/style-dictionary/#/README

fun + open source ~flow-tool
💾 https://wireflow.co/

• really digging the team/collaborators section that includes details like number of commits
• responsive minimaps are always appreciated
🧑‍💻 https://github.com/vanila-io/wireflow

just reopened notion && was greeted with a calendar announcement

💾 https://www.notion.so/product/calendar/download

poked around after reading equals' ai announcement && this scroll pattern is pretty fun (plus the domain is incredible)

🌐 https://equals.com/ai/
🌐 https://equals.com/command-k/

just noting the current example prompt patterns in bard v google ai studio.

was fun to see it auto-refresh the first time i opened it + update based on model
(tab to apply feels pretty good in a more ~dev-env, curious how these patterns will evolve)
https://makersuite.google.com

i had a half-waking nightmare i was the host of figbrew for a moment.
note: if someone asks you to not pop-quiz them, please don’t proceed with the exam.

(my mathing ability is ~somewhere around my cat's)

📼 https://youtu.be/3HhBpfyKH_M?si=TGcEKlMeVm9A3Wc5&t=1785

physical book just arrived 🎉

📖 https://rosenfeldmedia.com/books/design-that-scales/

absolutely wild site creator from mozilla.
there’s name suggestions, provides a url, && was only a couple clicks. 
(i'm still trying to process this experience)

🤖 https://soloist.ai

(peek the unedited default outputs)
🤷 https://soloist.ai/unwinddesign

i’m really liking the format of these interviews.

• wide range of guests + use cases.
• the questions get very specific on how* people are thinking about && using chatgpt, including sharing their screen.

📼 https://www.youtube.com/@EveryInc

some notes before morning dog walk routine
(just getting rough thoughts out on having an 'active' v 'passive' approach to ai+prompts && people confusing ai-output w/actual thinking, likely won't make sense)

Posts