Why even have product teams just let the agents implement whatever random requests customers tweet on here 😂
(I’m kidding)
Sorry, but this is wrong.
Check back in two years.
Project Syndicate: Something big may be happening for AI startup founders like @mattshumer_, but that doesn't mean we'll see it in the productivity statistics. @carlbfrey explains why. https://bit.ly/4u71o5h
Haha, uh, I guess one thing I would recommend to young founders is read a lot
Like, read *A LOT* just learn about the world. If you like to read it's not a waste of time. It might be the best use of time.
Products that need APIs/MCPs:
@substack
@RiversidedotFM
Every other video editing tool
Every bank that's not @mercury
Every government website
Every healthcare portal
how is gpt-5.5 performing for you?
VraserX e/acc: GPT-5.5 feels insane so far.
What I love most is that it is blazing fast, but still clearly smarter than GPT-5.4.
In my own tests, it feels around 2x to 5x faster depending on the task.
What’s your experience so far?
how is gpt-5.5 performing for you?
VraserX e/acc: GPT-5.5 feels insane so far.
What I love most is that it is blazing fast, but still clearly smarter than GPT-5.4.
In my own tests, it feels around 2x to 5x faster depending on the task.
What’s your experience so far?
I knew B2B marketing slop even when I was a summer intern at my first B2B startup job (didn't like it much tbh, total waste of a summer)
Soleio
New season incoming.
Follow @firstofkind or sign up for an email when Episode 1 drops:
@rauchg of @vercel
First of Kind: $ firstofkind s3
loading…
Everything Chris Tate makes is 🔥
Chris Tate: Portless killed :3000
Dev servers got stable names like myapp.localhost
Agents could use worktrees in parallel without stepping on each other
Now it's easier than ever in v0.11
Just run: portless
Zero config. Zero args. Zero code changes.
Trip Gabriel
The EPA's independent science arm did groundbreaking research to save lives. It studied fertility, asthma, wildfires, drinking water, climate change etc etc
In just one year, it has been almost completely dismantled
Of 1500 scientists, only 124 remain
https://www.nytimes.com/2026/04/27/climate/epa-science-trump-cuts.html?unlocked_article_code=1.eVA.yCwJ.TYSUUslTTFoa&smid=tw-share
TheValueist
. @garrytan GBrain has 10x’d (if that's even possible) my belief in the power of agentic GAI. The agents are significantly more productive, capable, and long-lasting when supported by a proper database-driven project board. I (my @openclaw) have made numerous incremental improvements that have further increased productivity. I can’t recommend Gary’s GBrain or something similar enough.
$NVDA $MU $SNDK $LITE
TheValueist: . @garrytan has written a significant amount about how to implement and enhance @openclaw . Some of it is good, some of it is fluff. His G-Brain tool is outstanding, and I highly recommend it for anyone using @openclaw . It is essentially a hierarchical to-do list on steroids
productive happy hour with a number of openai engineering managers today
productive happy hour with a number of openai engineering managers today
Jealousy 尼卡
我从几千个 Hermes Skill 里面精选了5个你必装的Skill
大家注意，如果你的Hermes不装Skill就不是完全体，我个人感觉最少差60%以上的能力。现在Skill仓库估计有几千个了，很多人看的眼花缭乱，我给大家精选了下面5个Skill，人人都用得到，人人都必须按照。
1. https://github.com/garrytan/gstack 85K星的含金量
必备Skill，优化Hermes记忆层，自我进化的必备Skill
2. https://github.com/garrytan/gbrain 11K星
同样是gbrain大佬的作品，多团队协作，效率提升的必备Skill
3. https://github.com/nesquena/hermes-webui 4.6K星
WebUi，轻量级设，让你的Hermes访问起来更加丝滑
4. https://github.com/0xNyk/awesome-hermes-agent
1.9K 星
全网最全的Hermes生态导航，必装的搜索级Skill
5. https://github.com/NousResearch/hermes-agent-self-evolution 2.4K星
Hermes创始人 @Teknium 亲自编写，自进化增强版。
上面5个Skill 可以说是必装的，我是尼卡，如果你也对 AI，Web3感兴趣，请关注我，给你分享更多好工具好项目。
Jealousy 尼卡: 橙皮书：Hermes Agent 从入门到精通
很多朋友不知道 Hermes Agent是什么？ 或者需要进阶攻略。现在好了，知名大佬花叔出了一篇 Hermes Agent 橙皮书，而且有原生中文版。
GitHub 上短短几天已经飙升至3K星了。
这本书我愿意称为是 Hermes Agent
Tibo
We will ship again this week. Codex has achieved escape velocity and will keep improving rapidly.
在悉尼和稀泥
今天 AI 圈最有意思的不是又一个新模型，而是 .@garrytan 公开了他个人 agent 的三个配置文件：SOUL.md（性格）、AGENTS.md（工作流）、USER.md（对象档案）。
结果？"从不会说话的工具变成了凌晨 2 点你愿意聊天的朋友。" 🧵👇
shipping for our users
Tibo: We will ship again this week. Codex has achieved escape velocity and will keep improving rapidly.
shipping for our users
Tibo: We will ship again this week. Codex has achieved escape velocity and will keep improving rapidly.
Hey
Honestly impressed with how simple @tibo_maker makes it sound but of course this comes from his experience.
Most people (me included) should just follow the core principals and not over think our own approach.
I'm following suit for #1 and #3. Can't quite do #2 yet cause not enough users.
Peter Yang: My top 5 takeaways from @tibo_maker, a solo AI founder who's making $1M+ a month:
1. Charge money on day one.
Tibo’s first startup failed because he cared more about appearing successful (e.g., I managed a team of 10 and raised $200K) than validating demand with paying
Everybody eats. Improving our food supply and safer more efficient agriculture is one of tue most direct ways for AI to help people
Calley Means: @ycombinator @garrytan Thank you for the leadership @garrytan in spurring leadership in technology for our farmers - who are the key to reversing our skyrocketing rates of chronic disease.
Frank Wang
if @garrytan can build gstack and run YC, there's no reason an engineering leader shouldn't write code, especially at a company trying to use more AI.
gotta be the change you want to see
Tibo
Don't just reset Codex rate limits for fun, it costs money.
Don't just reset Codex rate limits for fun, it costs money.
... but the vibes are good ...
I have reset Codex rate limits for ALL paid plans to celebrate a good week and allow everyone to build more with GPT-5.5. Enjoy
Soumya Gupta
For a little over a month I've been going nuts with slash commands, thanks to @garrytan's Gstack hooked into @conductor_build.
Yesterday, I uncovered a new joy - putting together a product deck to share with my clients. Claude Code picked up my product branding, took copy inputs from me and codex, went through multiple iterations - you can use /design-review and /design-html on marketing materials too, and made me very happy. All in 3 hours of conversation.
Building product was always fun for me, but building the marketing materials? Not so much.
So pumped to have the coolest AI setup (that keeps getting better!) to help me build AND sell.
For the curious, here's the deck: https://telborg.com/decks/telborg-dc-intel.pdf
Marc Andreessen 🇺🇸
Actual anti-AI energy is the newest instance of @robkhenderson’s “luxury beliefs” — a small fringe of elites playing at radical politics. For normal people, AI is a useful app on their phone, and they’ll just use it more and more.
Marc Andreessen 🇺🇸: It won’t happen. People say all kinds of things in surveys. Watch what they do instead. And what they do is… use AI.
I guess it's pretty satisfying to update GBrain and have my OpenClaw say "That's the money shot right there" 👀
BTW new stability release on migration verification
Has anyone gotten clean, seamless clip-to-clip extends working in Seedance 2.0? How?
By seamless, I mean:
- no camera cuts
- no visual or audio stutter
The anti-tech lines are so tired, and the people who hate tech have hated tech for so long
So of course if it's the same people doing the same old dance, they reuse the same old lines when actually everything changed in 2026
Andy Masley: Jon Oliver repeats the line "chatbots are designed to maximize the time you spend on them" and "single-mindedly pursue human approval at the expense of all else." A lot of the quoted stories are from over a year ago.
I don't know how this is still getting said in 2026.
Fudge I'm here to help YOU get the bag
fudge: @almostpmf nah its okay i hope they get the bag
GStack ships with Chromium that has Claude Code as the CLI and you can type /pair-agent from your real CC or the sidebar one
And it'll give you a string to paste into your OpenClaw/Hermes so it can control your browser remotely, and you can log it into stuff and have it scrape
The artist is closer to reality than the CEO.
Here's a fun skill you can teach your OpenClaw/Hermes, just paste it and find a book and it'll give you hours of fun
Book Mirror — How It Works
What it does: Takes any book + everything the AI knows about the reader → produces a personalized chapter-by-chapter analysis. Left column: what the author says (detailed, preserving stories and frameworks). Right column: how it applies to this specific person (using their actual words, situations, people, dates from memory).
The pipeline:
1. Get the book — PDF or EPUB
2. Split into chapters — one text file per chapter
3. Build context pack — pull everything relevant about the reader: bio, patterns, recent conversations, relationships, active projects, emotional landscape
4. Per-chapter analysis — for each chapter, produce a two-column table:
| What the Author Says | How This Applies to You |
| -------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| [Preserve the actual stories, stats, quotes, frameworks — detailed enough you could skip the book] | [Name real people, dates, situations from the reader's life. Read like a therapist who knows them writing notes in the margins.] |
5. Fact-check — verify every personal claim against source material
6. Generate PDF — deliver
The quality bar:
• Left column: you shouldn't need to read the original book
• Right column: should feel like someone who actually knows you read it first
• If a chapter doesn't apply, say so — never force connections
• No generic advice ("consider reflecting on...") — only specific mirrors
What makes it work: The right column is only as good as the context. A book-mirror for a stranger is just a book summary. A book-mirror for someone whose memory system has years of conversations, therapy notes, relationship patterns, and exact quotes — that's a different thing entirely.
Example output: A philosophy book about ego and identity → mapped to the reader's specific shame patterns, family dynamics, work style, and therapy breakthroughs, using their own words back at them.
Timo Heuer
Finally found some time to play around with @garrytan's Gstack (https://github.com/garrytan/gstack) in combination with @conductor_build.
Incredibly strong setup and Gstack already feels like a powerful resource for building 🏗️
GPT-5.5 can follow your requested response style
Simon Smith: GPT-5.5 is the first model in ChatGPT to which I’ve been able to just describe my desired response style, have it understand and remember it, and then actually respond in that exact way.
I want it to write like a normal person, to default to narrative with some personality, to
GPT-5.5 can follow your requested response style
Simon Smith: GPT-5.5 is the first model in ChatGPT to which I’ve been able to just describe my desired response style, have it understand and remember it, and then actually respond in that exact way.
I want it to write like a normal person, to default to narrative with some personality, to
imagegen changing how product ideas are shared
Rohan Varma: Our 2.0 image model is so good at making screens and vision mocks.
Something about AI generated images of digital surfaces feels very “right” to me.
Internally, I’ve started seeing tons of product ideas shared and brought to life via image generation rather than prototyping —
imagegen changing how product ideas are shared
Rohan Varma: Our 2.0 image model is so good at making screens and vision mocks.
Something about AI generated images of digital surfaces feels very “right” to me.
Internally, I’ve started seeing tons of product ideas shared and brought to life via image generation rather than prototyping —
shipping velocity is high
Simon Smith: This is amazing. OpenAI is incredibly responsive to feedback. I noticed the 360° image trend after GPT Image 2, commented about it to @JustinBleuel on April 22, and we got a new feature five days later from @adele__li. Now let's see what virtual worlds people build and share!
shipping velocity is high
Simon Smith: This is amazing. OpenAI is incredibly responsive to feedback. I noticed the 360° image trend after GPT Image 2, commented about it to @JustinBleuel on April 22, and we got a new feature five days later from @adele__li. Now let's see what virtual worlds people build and share!
Solo AI founders tend to get stuck trying to make one product work. Here's @tibo_maker advice on what you should do instead:
"We had this crazy phase where we were shipping one new product a week until something stuck.
We were just printing the landing page, tweeting about it, and if there's no traction, we go to the next product.
Nine products failed that way and the 10th one just took off.
I would not have bet on my successful products in advance. They were a surprise to me too.
I think you need to change your mindset and see every product release as an experiment. It can totally fail, you just go to the next experiment and continue."
tl:dr Ship early and often, don't try to make it perfect.
📌 My full interview with Tibo is full of practical advice for bootstrapping AI products: https://www.youtube.com/watch?v=0UnZnonMN9o
Peter Yang: "I shipped 9 failed products before one took off...now I'm doing $1M+/month."
Here's my new episode with @tibo_maker, a solo founder who bootstrapped 5 AI products to $1M+ / month.
Tibo walked me through his exact playbook:
✅ How to validate ideas and fail fast
✅ Why his top
Peter Yang
Solo AI founders tend to get stuck trying to make one product work. Here's @tibo_maker advice on what you should do instead:
"We had this crazy phase where we were shipping one new product a week until something stuck.
We were just printing the landing page, tweeting about it, and if there's no traction, we go to the next product.
Nine products failed that way and the 10th one just took off.
I would not have bet on my successful products in advance. They were a surprise to me too.
I think you need to change your mindset and see every product release as an experiment. It can totally fail, you just go to the next experiment and continue."
tl:dr Ship early and often, don't try to make it perfect.
📌 My full interview with Tibo is full of practical advice for bootstrapping AI products: https://www.youtube.com/watch?v=0UnZnonMN9o
Peter Yang: "I shipped 9 failed products before one took off...now I'm doing $1M+/month."
Here's my new episode with @tibo_maker, a solo founder who bootstrapped 5 AI products to $1M+ / month.
Tibo walked me through his exact playbook:
✅ How to validate ideas and fail fast
✅ Why his top
Latent.Space
The $15B Physical AI Company: 4 stack rewrites, end-to-end RL, neural sim, android for vehicles, trucks in Japan https://latent.space/p/appliedintuition
@AppliedInt co-founders @qasar (CEO) and Peter Ludwig (CTO) explain why physical machines today look like phones before Android, why the real bottleneck in physical AI is deployment on constrained hardware, how neural simulation and statistical validation make autonomy safer, why Applied runs L4 driverless trucks in Japan, and why after 10 years they can predict the next 20 problems every robotics demo will hit.
jack
Garry Tan: The artist is closer to reality than the CEO.
GBrain ecosystem is 🛫
Gopi Vikranth: A comment I received recently is that agents cannot really do analytics and data science yet. It is still mostly hype/slop. I disagree. That may have been true a year ago. I do not think it is true now. LLMs have crossed an important threshold.
The question is no longer whether
Charles E Gaudet II
Buckminster Fuller discovered that the real story behind Darwin's "Survival of the Fittest" has to do with helping others 'survive' by delivering the most value.
The more you help others do more, get more, etc in an authentic way - they conspire around you to help you succeed.
It appears @garrytan has figured this out. He continues to give HUGE value to the AI community in a massive way.
And the community continues to cheer him on.
Garry Tan: GBrain is at Apple I stage: you still have to solder your own breadboard and it's a motherboard, not a full computer... but it won't stay like that forever
https://github.com/garrytan/gbrain
Manny Bernabe
We're live! Going deep on Security at Replit w/ CTO @lhchavez !
GET IN HERE!
https://x.com/i/broadcasts/1YxNrZYVeoZxw
Firecrawl
Introducing our /parse endpoint 🔥
Convert any document into clean data for AI agents via API.
- Turn PDF, DOCX, and XLSX files into Markdown or JSON
- Preserves reading order and tables
- Zero Data Retention support
All powered by our 5x faster Rust-based parser.
Replit ⠕
You're going to be embarrassed by the slides you made before AI
Meet Replit Slides
The first AI slides with stunning design
Samuel Spitz
There are countless AI slide products
But we decided to make another one
Why? We made an architecture breakthrough allowing us to write PPTX-compatible code in a completely different way than competitors.
This unlocks much prettier slides. See the real examples below!
Replit ⠕: You're going to be embarrassed by the slides you made before AI
Meet Replit Slides
The first AI slides with stunning design
The slides are actually really good. I used to pay people to make me slides and the process was always frustrating, but now I actually enjoy making them.
Replit ⠕: You're going to be embarrassed by the slides you made before AI
Meet Replit Slides
The first AI slides with stunning design
Zhen Li
Most AI slide tools still look like templates
You can tell in 2 seconds it’s AI
We made Replit Slides because of that
It's built to design slides, not just fill them
Replit ⠕: You're going to be embarrassed by the slides you made before AI
Meet Replit Slides
The first AI slides with stunning design
Kane 謝凱堯
San Francisco public schools @SFUnified (53% reading, 42% math proficiency), in the midst of a budget crisis, paid for a workshop on “Adultism” and “Adult Supremacy”.
I’m going to request the public records on how much they paid for this.
Rachel Keuler: Adults are “oppressors” and kids should stop listening to their parents and teachers.
Is the hope that we will convince kids that adults are bad…except for “Activist Adults”?
Browser Use
APIs are dead
Johannes Dittrich: I automated my image generation via Midjourney with Browser Harness..
Midjourney has no API, so I asked Claude Code to use Browser Harness to automate it:
> Wrote prompts
> Used Midjourney UI to create images
> Created tools to make later runs faster
Try it now
Samuel Spitz
Talking about this launch now with the OG creator of Replit Agent @zhenthebuilder and @biknandy who built slides with me!
https://x.com/i/spaces/1AKEmOMjwPVKL
Replit ⠕: You're going to be embarrassed by the slides you made before AI
Meet Replit Slides
The first AI slides with stunning design
Replit ⠕
Join us for spaces on Replit slides now
https://x.com/i/spaces/1AKEmOMjwPVKL/peek?s=20
Peter Hague
No system in human history has allowed more people to sit in their garden making art than capitalism
𐌁𐌉Ᏽ 𐌕𐌉𐌌𐌉: I just really, really hate the concept of capitalism. I don’t want to work my job every day. I want to make art. I want to be in my garden. I want to feel something.
Sheel Mohnot
Everyone is getting this wrong:
The politicians in California are pretty terrible, but you can't blame them for the wealth tax... they actually killed the idea 2 years ago.
Our uniquely bad form of direct democracy is the actual culprit.
Democratic Socialist assemblyman Alex Lee introduced the wealth tax in Sacramento (AB 259) 2 years ago. It was not popular, it actually did not even make it to a vote. It was deemed too leftist even for the California legislature.
So why is it back?
California's uniquely bad form of direct democracy lets you bypass the legislature if you get enough signatures... You can put almost anything on the ballot if you get signatures.
That is what the SEIU-UHW, the healthcare workers union did. Union leadership spearheaded this initiative and funded the campaign directly to collect signatures for the wealth tax. They collected signatures by going around asking if people want more money from billionaires to fund hospitals, healthcare, food aid, and schools... naturally people said yes without realizing the consequences.
The only major currently elected official from California that supports it is Ro Khanna. My guess from seeing his support on twitter is that Ro decided to support it without understanding it and then dug his heels in for some stupid reason (he actually acknowledged that it is bad as written).
Tom Steyer and Saikat Chakrabarti who are running for office also support it (we need to do everything we can to oppose them).
I can't believe I'm defending California politicians, but remember - they actually rejected a wealth tax. This is a union-backed ballot initiative trying to go around them.
Kyle Tibbitts: Room temp IQ politicians in California looked at the greatest entrepreneurial ecosystem in the world and said “fuck it, let’s blow it up”.
god candle
on May 2nd, @replit agents are free to use all day.
free tokens. so zero excuses.
use the next few days to plan what you’re building, map the idea, outline the features, and know exactly what you’re launching.
then when the clock hits 12:00, get after it.
godspeed.
We're growing the Vercel Labs team. Our mission is to build the devtools of the AI era.
We used to build tools for humans, now we're building them for agents. @ctatedev & team have shipped the tools that agents love: 𝚊𝚐𝚎𝚗𝚝-𝚋𝚛𝚘𝚠𝚜𝚎𝚛, 𝚙𝚘𝚛𝚝𝚕𝚎𝚜𝚜, 𝚜𝚔𝚒𝚕𝚕𝚜, 𝚌𝚑𝚊𝚝, 𝚓𝚞𝚜𝚝-𝚋𝚊𝚜𝚑, 𝚓𝚜𝚘𝚗-𝚛𝚎𝚗𝚍𝚎𝚛… 22.8M+ downloads to date.
Dream job. DMs open 😀
Chris Tate: Portless killed :3000
Dev servers got stable names like myapp.localhost
Agents could use worktrees in parallel without stepping on each other
Now it's easier than ever in v0.11
Just run: portless
Zero config. Zero args. Zero code changes.
Google for Startups
When @GoogleDeepMind’s @DemisHassabis and @YCombinator’s @GarryTan sit down, it’s worth paying attention. 📝👀
They discussed exploring the path to AGI, the future of @GeminiApp, how startups can find alpha through atoms, and so much more.
Conversation coming soon. Founders, you don’t want to miss this one. 🔔
Deli Board is the best sandwich in SF and the world @bettersoma @war24182236
Jessica Livingston
Today's Social Radars is an exciting one: Ron Conway talks publicly for the first time about the frantic, behind-the-scenes efforts that prevented the failure of Silicon Valley Bank from triggering a Depression-style financial panic.
https://pod.link/1677066062/episode/MmRjMWUwMmUtNWEwYi00OTY2LTg1YTctZTRmYmU3MjFlNjAz
Abe Murray
I ingested my entire ebook library into @garrytan's gbrain 🧠📚
I have a go-to challenge for AIs - whether they can help me find this book I vaguely remembered
tl;dr - my book system + gbrain got the job done
(its Cryptonomicon, btw)
What if we measure how good AI models are by what era of games it can one shot a single level for (or maybe build with 30 min of work):
Atari
NES
SNES, Sega Genesis
PlayStation, N64
PS2, Xbox, GameCube, Dreamcast
Xbox 360, PS3, Wii
PS4, Xbox One, Switch
PS5, Xbox Series X/S
I think we're still in the NES era?
MikeCJunior
Honestly @Replit has cooked on the slides. I’m impressed.
wow y'all love 5.5
we should think of something nice to do to celebrate!
Abdulrahman
Huge day for consultants!
Replit ⠕: You're going to be embarrassed by the slides you made before AI
Meet Replit Slides
The first AI slides with stunning design
Ayush Tripathi
Builders... this is a must watch: https://www.youtube.com/watch?v=zB1lpGNWVtg&t=1624s
@Replit 's journey and all the chaos in between
Arnav Sahu
Wow. The Google for agents is Exa 🫡
Exa: We're excited to partner with Google to offer Grounding With Exa inside of Gemini models!
Using Exa's agent-first search, Gemini models can now access billions of websites, technical docs, papers, people, companies, and more.
10^18🤝10^100
Basile Terver @ ICLR 2026
What a week at @iclr_conf in Rio 🇧🇷 !
As my first conference, it would have been hard to ask for a better experience 😍. I had the chance to present two papers:
- EB-JEPA https://github.com/facebookresearch/eb_jepa: Presented as an Oral 🏅 (top ~3% of papers) at the World Models workshop.
- Dual Perspectives on non-Contrastive SSL https://arxiv.org/abs/2507.01028: Presented as a poster.
-> I’m already working on two follow-ups to these, so stay tuned for more on that 😉
I was genuinely impressed by the quality of the discussions during the poster sessions. A huge thanks to everyone who stopped by our poster to debate the details, and to those who attended the talk 🙏
I’m leaving Rio with a lot of new research ideas and potential collaborations. This week really made my motivation to keep pushing papers and staying active in the community even stronger for the rest of my PhD 🚀
Another highlight was the socials, specifically the @amilabs mixer. It was exciting to see AMI’s first official public event go so well—it definitely felt like we put AMI “on the map” for a lot of people.
Thanks again to my co-authors 🙏:
- EB-JEPA: @randall_balestr, @megdrv, @DavidJFan, @garridoq_, @TusharNagarajan, @koustuvsinha, @kevinghstz , Mike Rabbat, @ylecun, @_amirbar
- Dual Perspectives on NC-SSL: Jean Ponce, @MichaelArbel, Martial Hebert
Replit ⠕
Ready to have some fun building this weekend?
This was a huge highlight for me
Was an incredible discussion to hear Demis speak about the future. He is building it.
Google for Startups: When @GoogleDeepMind’s @DemisHassabis and @YCombinator’s @GarryTan sit down, it’s worth paying attention. 📝👀
They discussed exploring the path to AGI, the future of @GeminiApp, how startups can find alpha through atoms, and so much more.
Conversation coming soon. Founders,
It's honestly impressive that GitHub kept the service up at all, given this kind of growth.
I predicted this years ago: Free services will become untenable with the advent of human-level bots.
Worth exploring micro-payments: Even cents per git push might be enough to reduce spam and make this sustainable. Maybe powered by Bitcoin to keep this open and accessible (as opposed to KYCing users).
Mitchell Hashimoto: Ghostty is leaving GitHub. I'm GitHub user 1299, joined Feb 2008. I've visited GitHub almost every single day for over 18 years. It's never been a question for me where I'd put my projects: always GitHub. I'm super sad to say this, but its time to go. https://mitchellh.com/writing/ghostty-leaving-github
Many such cases
Though also sorry it is still very experimental! It’s far from easy to use yet
Preetham Kyanam: @garrytan i'm really liking GBrain. Just got it setup with my Hermes agent. Having it store all sorts of information for me right now. I wonder what else it's capable of
Manny Bernabe
Replit Agent will be FREE for 24 hours on Saturday. 🤯🤯🤯
Replit ⠕: Free Agent usage starts at 5:00am PST on May 2 for 24 hours
Thrilled to support the education GOAT as he reinvents education (again).
Winston Brown
My current “take over the world” coding stack:
⁍ Hermes Agent
⁍ GBrain
⁍ GStack
⁍ Matt Pocock’s Skills repo
⁍ coffee
⁍ delusion
⁍ and terminals I’m afraid to close
Agentic coding got weird fast.
A year ago the stack was:
‣ React
‣ FastAPI
‣ Postgres
‣ Redis
‣ Docker
Now it’s:
○ agent kernel
○ memory layer
○ skills layer
○ browser tools
○ subagents
○ permission boundaries
○ human supervisor with emotional damage
@garrytan is giving agents a brain.
@mattpocock drops a Skills repo and suddenly my coding assistant needs onboarding docs and SOPs.
The shift isn’t “AI writes code.”
It’s that the dev environment is becoming an operating system for tiny software employees.
Are we still coding, or founding micro-companies inside the terminal?
Manny Bernabe
BREAKING: @ViktorThulin just joined the presidential advisory group on AI in Iceland. 🇮🇸
From Replit builder to advising the President.
viktorTJM: A week ago, I had the honor of being invited to join the presidential advisory group on artificial intelligence.
A little while ago, I reached out to the President of Iceland, Halla Tómasdóttir, through a mutual connection to encourage greater public focus on AI within Iceland.
Jitendra MALIK
Re @jon_barron "World models" has a technical meaning - the transition model/dynamics model from Bellman/Kalman in the context of MDPs/ state space approach to control theory ~ 1960. I gave a talk on this history https://www.youtube.com/watch?v=9B4kkaGOozA&t=3891s
x.com