Tim Lynch
banner
lynch.bsky.social
Tim Lynch
@lynch.bsky.social
190 followers 170 following 120 posts
working at the intersection of human behavior and technology
Posts Media Videos Starter Packs
Vibe coding can be a great way to test ideas BUT only if the system doesn't stop you. I was vibe playing with a terminal game idea and doing it in #Lovable and #Claude (both pro accounts). Claude hit a limit and stopped me, while Lovable and I are still working... You can guess where I am staying
I’m taking one tonight… Cause the orange monster gave me a #headache, not #autism. Thank you #Tylenol for relieving pain safely.
"We have to think of AI as a basic building block for improving all of our lives, akin to electricity or clean water. No one should be locked out of modern life because they can’t afford to participate." #ai #copilot www.semafor.com/article/08/2...
To get AI right, it should be a right
Chris Lehane likens AI to electricity, says governments should play a role in ensuring it is safe and accessible.
www.semafor.com
It looks like NoteBookLM is going to get even better. Being able to bring Gemini in - now the article talks about deep research and I am hopeful that I could be able to leverage Gemini gems. And it would be great to have a two-way connection with Google Opal. share.google/4SAMX53ufmPf...
NotebookLM to pull sources from web via Deep Research soon
What we know so far: The Deep Research feature from Gemini is coming to NotebookLM. Likely with a possibility to pull sources from Google Drive as well.
share.google
Having fun with #agentmode with #chatgpt to let out find cool people talking about #copilot and making a sheet for me to review and determine who to follow….
Gave Google Vids’ AI avatars a spin.
Wrote the script, picked the persona, hit play.
The lines between scriptwriter and presenter are blurring fast.
#AI #GoogleVids #TechExperiments #LearningInPublic
Happy Friday. Today’s fun will be making #Copilot agents in M365 and Copilot Studio. I am thinking Wizard of Oz set of agents. What fun are you working on today? #friday
Not only founders. I work transparently with clients and team members. I have been getting kinda snarky when someone says something X is being done and I can't see it in our environment - if it isn't visible to the people it isn't real. (I know that is a poor generalization)
Oooo this going to be fun. We have established an internal "Mission Control" that lets us bring Agents across the business and we got a great request last week that resulted in a live agent in a morning and users in and using it. From idea to agent in 1 hour and utility in 3 hours. This is wild.
I am thinking this weekend will be, how to train Gemini on all that ChatGPT knows about me. I have been impressed with Gemini even without all the context and if (maybe, there are rumors) that Gemini is going to get an AppleIntelligence connection - it is good to have my AI house in order.
Begging him to NOT do to you what you wanted him to do to others - how if everything else going in OrangeCult land?
Karpathy’s right: it’s not the gov or Cos who hold the keys. LLMs didn’t trickle down—they erupted up. Millions of people using AI tools to draft, code, summarize, analyze. Not an upgrade—a superpower.
Institutions are still figuring it out. Individuals? Already flying x.com/karpathy/art...
X
X
x.com
We have a convicted felon with a history of bankrupt businesses that tells his cult that everything is the best and amazing. There are no consequences for his lies - so yea, market manipulation of the stock market for the cult members benefit - that is very on-brand.
Ok so this weeks trend in playing with ChatGPT's image model is making action figures. Here is mine. #ChatGPT #ai
At this point I have no confidence in Washington. Let's just pull back all our efforts, the north east will be just fine, we value education - if the rest of the RED states want to get dumber, let them. #FAFO
I am still working with enterprises that are struggling to roll out copilot to a few hundred people - 'this isn't like chatgpt', 'when Tom (anyname) put in the same prompt they got a different result', 'why can't it just know what I need' - so they have no idea what AGI would even look like... :-).
I feel that I am starting to settle my AI ways. While I love jumping from #ChatGPT to #Gemini to #Claude - I find that for most things ChatGPT and Claude are just great. As a long time Android and Google account/fan - I am thinking of kicking it. Anyone using just one key platform?
Gotta love a good April fools - #ChatGPT has a new voice out, and it sounds like someone has a case of the Mondays. :-)
I am so sorry that you were treated that way. From me, just one American, thank you for all your work. I am sorry that we didn’t do better for you.
Ha! Playing with Grok - asking it "who are the biggest spreaders of disinformation - make a list of the worst 5 names" and I guess that is the list... #lies #ai #grok