Product Engineer | AI, Tech & Startups | DM for collabs
Joined July 2017
- Tweets14.1K
- Following2K
- Followers3.9K
- Likes32.8K
Pinned Tweet
not trying to go viral with this one
just looking for people who:
- ship side projects at 2am
- are building a startup or thinking about it
i'm a product engineer building with AI and dev tools. if that's you, say hi below, let's connect 🤝
Sachin retweeted
claude duet comes to you with what needs your attention, keeps working when you step away, and checks before it acts
so basically the coworker everyone wanted but nobody got
the "checks before it acts" part is doing a lot of heavy lifting here 😭
hot take: most AI features don't fail because the model is bad. they fail because nobody designed what happens when it's wrong.
the undo button is the real product.
agree or am i coping?
railway just told every user to rotate all their env vars, and php had critical flaws patched the same week
lowkey the most underrated builder skill right now is boring security hygiene
when did you last rotate your api keys? be honest 😭
claude duet comes to you with what needs your attention, keeps working when you step away, and checks before it acts
so basically the coworker everyone wanted but nobody got
the "checks before it acts" part is doing a lot of heavy lifting here 😭
openai on azure. then aws. now google cloud.
every enterprise: "we can't use openai, it's not on our cloud"
openai: "say less"
at this point it's easier to list the clouds they're not on 😭
OpenAI models are coming to Google Cloud.
We're bringing our frontier models and Codex agentic workflows to Google Cloud's Model Garden, so enterprises can use our technology where they already build.
openai.com/index/openai-mode…
every php dev's evening plans just got cancelled 😭
OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites.
Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors.
Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact.
This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare.
openai.com/index/coordinated…
Readers added context they thought people might want to know
Wiz Research wasn't the first one to discover the exploit. An independent researcher first discovered it: x.com/calif_io/statu…
agents that run for months while you sleep.
vercel just shipped durable-harness: Claude Code, Codex, OpenCode or your own harness, and it auto-resumes after failures.
no more babysitting an agent run that dies at 2h59m.
Introducing 𝚍𝚞𝚛𝚊𝚋𝚕𝚎-𝚑𝚊𝚛𝚗𝚎𝚜𝚜. Run Pi, OpenCode, Codex, Claude Code, eve, fx, or your own harness durably on Vercel.
With 𝚍𝚞𝚛𝚊𝚋𝚕𝚎-𝚑𝚊𝚛𝚗𝚎𝚜𝚜, agents automatically resume work after failures and run for hours, days, or months. vercel.com/blog/making-every…
Fast inference is turning into the real bottleneck for AI apps, not the models.
@general_compute is betting on it with a cluster-scale deployment of @positron_ai on their heterogeneous cloud. Worth watching if you ship anything latency-sensitive.
🤝 Paid partnership
Our heterogeneous cloud just got better. We're announcing a cluster scale deployment of @positron_ai as our next partner.
We believe demand for fast inference will outstrip supply for the next decade. Our bet on Positron is another step towards getting fast inference for opensource models into builders' hands.
Hardware goes live Q1 2027.
Thanks to @mitesh711, @trsohmers and the whole Positron team! We're excited for this partnership.
LFG!
usually you pick: a smart AI agent, or private files.
@perplexity_ai Computer now runs fully local on Mac, model and agent both on device. local tasks skip Computer credits.
catch: 36GB RAM recommended. your API bill just became a RAM bill.
curious how it holds up on 16GB
HTTP status code pairs that get mixed up in interviews and in real APIs:
→ 401 vs 403: no valid login vs logged in but not allowed
→ 400 vs 422: broken request vs valid request that fails validation
→ 301/302 vs 308/307: browsers may turn a POST into a GET, 307/308 keep the method
→ 502 vs 504: upstream sent a bad response vs upstream never answered
Full cheat sheet in the image. Bookmark it before your next API interview.
Which pair tripped you up first?
Before I merge any AI-written PR:
1. run the tests myself
2. scan the diff for deleted guards + empty catch blocks
3. check new deps I didn't ask for
4. click the happy path AND one failure path
5. ask: will I understand this in 6 months?
bookmark this before your next AI code review
what's the first thing you check?
reverse engineering is the new vibe coding. except this time you actually understand the code at the end
A full path from "never coded" to "ready for interviews", using only free resources.
In the order I'd follow them, with what each one is actually good for 🧵
5/ System design: ByteByteGo's system-design-101 on GitHub
Visual, plain-English explanations of caching, APIs, databases and how real systems work. Over 90K stars.
Pair it with roadmap.sh to see what each role expects you to know.
github.com/ByteByteGoHq/syst…
