Building @tryskilly — Voice AI that watches your screen and moves your cursor to teach you Open source • Embeddable for SaaS Indie hacking from Alexandria, 🇪🇬
Alexandria, Egypt
Joined February 2026
- Tweets1K
- Following2.5K
- Followers375
- Likes645
Mohamed Saleh retweeted
Anthropic -> OpenAI, second one today.
lots of news breaking these days so I may as well sneak mine in if anyone cares and share that I recently joined @OpenAI 🫡
my time at Anthropic was a fever dream in the best way possible. got to help build the growth function, work alongside some really really talented people, and tackle some of the most challenging problems of my career.
incredibly grateful for the experience and excited for what’s ahead. also in case you were wondering, Ant has better coffee/drinks and OAI has better snacks.
Mohamed Saleh retweeted
welcome @_catwu to the team!!
Some news: I’m joining OpenAI to lead product management for Core Products!
I have a lot of respect for OpenAI for showing the world that scaling works, for bringing AI to over a billion people, and for believing great products are core to the mission.
I’m excited to work with @thsottiaux, an incredible PM team, and everyone across OpenAI. Together, we’ll build agents that have your back and your team’s, and keep you in control, so you get more time for the people and work you love.
Lots to learn, and lots to build. Please keep your feedback coming!
Mohamed Saleh retweeted
Canva is laying off 22% of its workforce, but the way it’s handling the layoffs deserves real respect.
Here’s what employees are getting:
- 6 months of salary
- 12 months of healthcare
- Career transition support
- $10K to start their own business
- Accelerated equity vesting
- Can keep their MacBooks
CEOs everywhere should take notes.
Mohamed Saleh retweeted
update your VPS asap
AWWWWOOOOOOOGAAAAAAA
9.8 severity use-after-free vulnerability in the kernel’s IPv4 TCP stack just landed
Time to patch!
ubuntu.com/security/CVE-2026…
Mohamed Saleh retweeted
Person who is reverse engineering his tv gets banned from OpenAI
You aren't allowed to stop your tv from spying on your family. You have to ask for permission
Or, maybe, maybe we accelerate
Mohamed Saleh retweeted
Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index scoring 55, making large gains in agentic performance; it will be the leading open-weights model once its weights are published
@Kimi_Moonshot's new reasoning model gains 11 points over Kimi K3 (max, 44). It trails only Claude Opus 5.5 (max with fallback, 58), Claude Sonnet 5.5 (max with fallback, 56), and Muse Wonder Full (max, 56) on the Intelligence Index.
Once the weights for K3.1 Pro are released it will be the highest-scoring open weights model and the gap between the leading open weights and proprietary models on the Intelligence Index will narrow from 12 points to 3.
Key results:
➤ Large gains in agentic performance: Kimi K3.1 Pro scores 3rd on GDPval-AA with 1,846 Elo, up 312 Elo from Kimi K3, behind only Claude Opus 5.5 (1,866 Elo) and Muse Wonder Full (1,857 Elo). Kimi K3.1 Pro also achieves the highest score of any model on Terminal-Bench 4.0. K3.1 Pro scores 67%, up from 13% for Kimi K3 and ahead of Claude Sonnet 5.5 (64%). AutomationBench-AA, which tests agentic SaaS workflows, rises from 58% to 73%.
➤ Gains in other evals are more moderate: Humanity's Last Exam rises from 47% to 52%, CritPt from 23% to 25%, and SciCode from 59% to 61%. AA-Omniscience Index rises from 20 to 30, as the hallucination rate falls from 53% to 44% and accuracy rises from 48% to 51%.
➤ Cost per Task roughly equal to Kimi K3 at $1.98 per Task: K3.1 Pro uses 55K output tokens per Intelligence Index task, 14% more than Kimi K3 at 48K, while its first-party token prices are 33% lower.
Additional model details:
➤ Context window: 1M tokens.
➤ Size: 2.8T total, 104B active parameters
➤ Input modalities: Text and image
➤ Weights and license: Weights currently not published, Moonshot is expected to release the weights soon after launch
➤ Pricing: $2/$10/$0.20 per 1M input/output/cache hit tokens, 33% lower than Kimi K3 on all three rates.
Congratulations to the team at @Kimi_Moonshot on the release!
Mohamed Saleh retweeted
We recently hired a super CRACKED infra engineer!
People always ask me what it's like to 'manage' someone with more experience than you...
Tbh it was uncomfortable at first until I understood these things:
- I'm here to guide him in the right direction, not micromanage how he does things
- We own different parts of the product and have complementary skills. I don't necessarily have to be better than him at everything to 'manage' him
- 'Manage' is not the right word to describe our dynamic. It implies one is superior to the other, when in fact we just have different strengths
After understanding these things, I no longer feel imposter syndrome.
We've been shipping faster than ever since having him on the team and will properly introduce him soon 👀
We just hired one of the strongest infra engineers I've met.
10+ years of experience building distributed databases, principal-level engineering, and deep key-value storage expertise.
In his first week at @insforge, he already built a very exciting remote coding-agent experience (stay tuned this week 👀)
Really excited to work with him, and I can feel our team getting stronger and stronger.
Will introduce him properly soon!
Mohamed Saleh retweeted
Software is having its Uber moment.
For years, SaaS companies could keep raising prices because building an alternative was too expensive.
AI changes that.
Now anyone can enter the market with a competitive pricing. And customers can vibe code a good-enough version themselves.
It's a bit fascinating watching what feels like panic in public from Tyler.
I'm in some groups with lots of newsletter operators. Many people are fed up with Beehiiv and their price hikes.
They are switching providers, building their own, or jumping to @marclou's that he built in like 3 hours.
Tyler is doing what we call "rolling out the red carpet on the way out".
He is now asking what he can do for customers after he has jacked the price up and they are leaving.
That's the opposite of what you should be doing.
The issue I keep hearing from so many customers is that there are a ton of bells and whistles and garbage that they have no interest in using but feel like they have no choice but to pay for.
It's like your TV provider telling you how you have access to 673 channels but you're like, "that's cool... how do I get a package with the 4 I care about?".
I don't have a dog in this fight because it's too easy to simply vibe code up all the parts of Beehiiv that I'd actually need and pay a fraction of the price.
But it sure is wild to watch.
Mohamed Saleh retweeted
DeepSeek might as well open source the whole company at this point.
“We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.”
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
Mohamed Saleh retweeted
China just put out a draft that could change how all of us use open-weight models.
Exporting weights, checkpoints or code of the very largest models (2T+ parameters, 10^26+ FLOPs, or top 10 globally) would need government approval. API and app access stays untouched so far.
If it goes through, the biggest Chinese labs may not be able to freely release their top models anymore, and teams in the US and Europe that build on them could feel it fast.
My guess is this is a response to the US and its chip controls.
Mohamed Saleh retweeted
Why so many paying Claude Pro & Max ($200) developers are waking up to the “Your account will remain closed” screen right now:
Anthropic banned 11.4M accounts in H1 2026 (up 7.9×), and 89.45% of appeals (356K of 398K) get permanently rejected.
The 5 automated triggers catching normal developers:
1️⃣ Third-party OAuth → Logging into OpenCode, Roo Code, Cline, or OpenClaw with a Pro/Max subscription instead of an API key
2️⃣ VPN & IP hops → Switching VPN nodes, using Hetzner/VPS datacenter IPs, or traveling abroad
3️⃣ Billing changes → Upgrading 5x → 20x Max or switching cards (GitHub Issue #5088)
4️⃣ Automated cron/bots → Calling consumer plans via scripts
5️⃣ Security prompts → Running cyber/red-team code without enterprise clearance
Export your local `.md` context regularly once this screen hits, there is no second appeal.
Mohamed Saleh retweeted
🚨 Tibo, STOP talking about plans. START executing.
We know your plan is to bring 5hr limits back and market it as: Users will get 2X usage in return for it.
But before that FIX CODEX.
We need 3 things:
→ A model that can actually beat Opus 5.5.
→ A model that doesn't burn through our quota like crazy.
→better usage limits. Give us predictable limits.
WHAT'S THE POINT of more usage if we keep hitting 5hr limits while the model is still weaker, slow and burns through our quota anyway? 😭
Mohamed Saleh retweeted
Turns out a lot of people wanted a Mac image editor with no subscription.
75,000 downloads in 3 weeks. Thank you!
robbietilton.com/compositor
I built a Photoshop-like image editor. It’s called Compositor.
robbietilton.com/compositor
I originally built it for myself to get off my Adobe subscription, but decided to release it for free and make it open source. It has all the essential tools I need for compositing, with none of the BS.
I know this workflow is kind of archaic in 2026, but I’m still using it until AI can actually get pixel-perfect on some of the details.
The entire app is 12MB. Photoshop is 6,455MB on my machine.
Mohamed Saleh retweeted
🚨Breaking: OpenAI and Anthropic executives are reportedly privately preparing for catastrophic AI events
According to Axios, they're planning for the fallout of a major AI cyberattack that could potentially take down banking, internet or power
Absolutely insane that we're already at this point, many AI industry insiders apparently believe a major incident will happen within the next 6 to 12 months
Mohamed Saleh retweeted
If you want to be the first to know about Grok @Bot updates, unique use cases, and pro tips, follow these accounts:
@poteto (Lauren) = Grok Bot lead engineer plus regularly roasts Tibo
@mattyp = drops digestible vid updates
@pengzheng_ = Grok Bot lead designer
@johnbai = designs Grok Bot
@lingxi = engineering on Grok Bot
@shaoruu + @baltaaazr = built the early foundation
@SamSokolin = connectors + Android
@jediahkatz = behind building the agent
@shubgaur = bot whisperer, and gave the founders talk
@ericzakariasson = release notes, tips + articles
@romanugarte_ = incubated, launched, and now leads product on Grok Bot
@joshkim = growth + creator program
@vincentzhu = how he uses Bot day to day
@kiaraplds = real use-case threads
@roshan_s = cohosted the Galaxy livestream
@kristaletz = enterprise GTM, runs the Grok Bot GTM community
@benln = Grok Bot communities, meetups + ship recaps
Bookmark this. They’re the reason why the Grok Bot changelog is shipping at an insane pace. Follow them and learn.
What's new in Grok Bot
0:00 Main @Bot
1:32 Engineering updates
2:18 Knowledge work and Google connectors
3:00 Team Bots in Slack
4:15 Team Bot voice calls and status lines
4:39 Plugin search with Command K
Mohamed Saleh retweeted
I got some issues with this.
The $132m/year is not real at all. Closer to $3m at today's prices, and as low as $1.2k within a year.
Yes, $132m -> $1.2k. Hear me out.
1. The price listed is not accurate at all.
$11m/month is just wrong. Real "api cost" is prolly $100k-$500k per month range.
The 1.5T tokens number includes cache'd tokens. I looked at my own usage, and 33.5B tokens came out to $0.56/m (under 1/10th of Peter's $8/m estimate).
I'm primarily using expensive models like Opus and Astra, and "1.5t tokens" would still have cost me ~$800k. If I used cheaper models and hit cache more, easily could get as low as $100k.
2. Intelligence is getting cheaper incredibly quickly
I know our LLM bills are going up and everyone is stressed about it. Totally reasonable! It's not because intelligence got more expensive.
Artificial Analysis tracks how "cost at a given intelligence level" drops over time. It used to drop at ~10x/year. Now it's at 30x lower in 5 months. (see image)
Assuming the $100k from before, cost could get as low as $3k/month in a few months, and if trends continue, $100/month by this time next year.
Obviously there will be smarter models and we will move to them, but replicating her experience today with equivalent models will be $1.2k/year by next year.
3. Nobody said this is normal
Lauren is the 0.0001% of "agentic engineer". Measuring her work as "$ in -> software out" is reductive and wrong.
Her spend should be looked at as a combination of product/eng, research, FDE work (pstack), and marketing. Worth $350k/day? No. Worth a mil or two a month? Absolutely.
She is exploring what happens when you ignore the cost of intelligence. She's not doing this because Cursor expects every company to spend $10m/year on tokens.
She's doing it because that $10m in tokens is going to cost $10,000 in a year or two.
poteto's cursor profile is public. She's used 1.5T tokens in the last 30 days.
Assuming an avg rate of $8/M that's $384k/day or $11m. In the last 30 days. Per year thats $132M.
Let's assume Cursor has some insane deal with Anthropic and gets 90% off. That comes out to $38k/day or $1.1M in Sept/Oct.
$13m will get me 26 sassy staff-level all stars for $500k/yr
Readers added context they thought people might want to know
The Cursor profile at cursor.com/@lauren shows 1.5T tokens used. Cost estimates at $8/M do not account for prompt caching common in agent coding, where cache reads are billed at ~10% of input rates per analyses of Anthropic pricing. This results in lower effective costs.
cursor.com/@lauren chudi.dev/blog/claude-pr… carlosaragon.online/blog/anthropic…
Mohamed Saleh retweeted
"Sir… Anthropic has officially destroyed our entire product lineup in one month… opus on high beat astra… opus on MEDIUM beat sol 6.1… we killed terra ourselves… sir… luna was our last hope… but haiku 5.5 just dropped and beat it on all 6 benchmarks… and sir… one more thing… tibo accidentally posted on main last night that our compute… is cooked…"
