ChaiBytesAI

1.3K posts

ChaiBytesAI banner
ChaiBytesAI

ChaiBytesAI

@chaibytesai

Building AI Customer Risk Engineer at https://t.co/LhBRp4qpyq, also built https://t.co/IckdHgf8uv. Prev @aws

Try Now 👉🏼 Katılım Haziran 2025
375 Takip Edilen175 Takipçiler
Sabitlenmiş Tweet
ChaiBytesAI
ChaiBytesAI@chaibytesai·
NEW DEMO One prompt. Six-sheet MBR. Two cents. "Generate a Monthly Business Review for February. Compare with January. Export as Excel." Recon spins up an @e2b sandbox, writes a Python script, queries the database, computes every metric, and builds the spreadsheet. No engineer involved. askrecon.com
English
2
2
12
1.4K
ChaiBytesAI
ChaiBytesAI@chaibytesai·
@RaphaelDabadie Me! Applying with askrecon.com - an AI Customer Success Engineer. Connects to internal systems like database, code, tickets, docs and monitor your customers and flags risks proactively for CS teams to focus on!
English
0
0
1
59
Hubert Thieblot
Hubert Thieblot@hthieblot·
You became a founder. You quit the 9-to-5. You raised a little money. Everyone called you "brave" over drinks. You spent your nights building and your days pitching "the future," convinced that the next launch would change your life forever. Then a year passes. Flatline traction. $0 salary. Your co-founder quit via Slack. Your girlfriend left for someone with a 401k and a "stable" future. Your friends are posting house keys while you’re staring at a bowl of ramen, rehearsing the same tired lies to your parents about why the "big break" is just around the corner. Is this the end? You start wondering if you made a mistake. No. You keep telling yourself every founders went through this at some points. But you don’t stop. Logic says quit. Your ego says run. But there’s a sickness in you that won't let go. You’d rather fail at this than succeed at anything else. You tell yourself it’s just one more launch, one more pivot, one more "yes." You’re not delusional; you’re committed. You’ll miss this. Not the stress, but the electricity. The raw doubt that forced you to grow. The quiet fire of building while the world slept. The pure, unrefined dopamine of that very first user. These aren't just "hard years", they are the years that forge you. One day, when the bank account is full but the mystery is gone, you’ll find yourself wishing you could feel this hungry again. They all do.
English
118
75
1.2K
45.5K
Jan Schnyder
Jan Schnyder@jjschnyder·
Doing @speedrun was one of the best decisions we ever made The team is awesome and helped us across sales, hiring and more. Very bullish on the program. 007 applications are open right now and the program only got better. You should probably apply :) P.S. Happy to give a referral + possible 10k scout check
Jan Schnyder tweet media
andrew chen@andrewchen

ok - dropping big dates/news for a16z speedrun: - starting TODAY, founders can apply for the 2026 program that runs July 27 to Oct 11 in SF here's the link: speedrun007.a16z.com/ac - we will be investing up to $1M and funding 70+ companies over the next few weeks - But there's also $5M in credits/tokens/etc from AWS/GCP/Open AI/Azure/NVIDIA/Deel/Stripe/etc. You'll also work with our amazing operating team (GTM, talent, brand, people, and more), and join our community of elite founders - we offer a Global Founders Program for international founders, to help with visas, banking support, relo recommendations - yes you can be solo (but better if you're further along, and have built a team). No you don't have to have an idea yet. Yes you have to know how to build (even if you're not technical) - Also, in other news: speedun is officially moving full-time to SF. (prev it alternated SF/LA) this is for all the obv reasons - we've continued to have an insane lineup of speakers, including the founders of Carta / DoorDash / Twilio / Figma / Zynga / Airtable / Twitch / and of course, lunch/dinners with Marc/Ben alongside a16z team - and much more - the deadline for applying is May 17!

English
51
4
127
13.8K
ChaiBytesAI
ChaiBytesAI@chaibytesai·
Solo founder. Spent 6 years at Amazon, on the founding teams of two 0-to-1 launches (Alexa Shopping, AWS SageMaker Unified Studio), then joined an AI startup last year as the 5th engineer, now 20. The Recon insight came from watching our PMs and CS team constantly pull engineers in to investigate customer issues, then realizing this is the shape of post-sales work at every B2B company. Building Recon in parallel today, going full-time the moment I land an accelerator or hit meaningful traction.
English
0
0
0
74
Natan Voitenkov
Natan Voitenkov@NVoitenkov·
Application for @speedrun 07 are officially open. The program not only changed @GenwayAI trajectory, but also ours as founders. I am looking for a few folks to mentor, way more valuable than a scout cheque if you ask me. If you haven't done so yet - comment with a blurb of what you are building, and I will reach out via DM. A tip I shared with folks that attended SR06: If you are a team of 3+ crashing in SF/LA for the duration of the program, don't be cheap! Get an apartment with more than 1 bathroom... otherwise it will turn into Sparta quicker than you think 😅
Natan Voitenkov tweet media
English
100
1
192
11.5K
Ishita Jindal
Ishita Jindal@IshitaJindal17·
Next @ycombinator application is due in 6 days. We announced that we got into YC P26 and lot of founders reached out asking what changed. So I wrote about it.
Ishita Jindal tweet mediaIshita Jindal tweet mediaIshita Jindal tweet media
English
17
7
189
15.6K
Eleftheria Batsou
Eleftheria Batsou@BatsouElef·
@chaibytesai You're right, team B wins, but as someone else said in the comments: "engagement bait gets impressions, authenticity gets clients", it depends on your priorities.
English
1
0
1
49
Eleftheria Batsou
Eleftheria Batsou@BatsouElef·
Be honest, which team are you? A. Authenticity B. Engagement Bait
Eleftheria Batsou tweet media
English
31
0
36
1.4K
Ishaan Sehgal
Ishaan Sehgal@ishaansehgal·
2am debugging a broken CI pipeline googled the error like a normal person top result: a GitHub issue that I had subscribed to four years ago @github please let workflows live in subfolders github.com/orgs/community…
English
1
0
10
387
ChaiBytesAI
ChaiBytesAI@chaibytesai·
Closed the loop on this. Spent the weekend going through every claim on the security page against actual code, and the exercise itself surfaced real improvements. One example: sandbox isolation. Used to be one sandbox per workspace. Now it's scoped to the workspace plus the exact set of connections, with per-conversation scratch directories so two investigations running at the same time can't see each other's notes. Useful resource to share with leads ahead of customer calls. askrecon.com/security
ChaiBytesAI@chaibytesai

Running Recon through Claude Powered SOC-2 security scrutiny yesterday. No external auditor, no cert, just me + Opus 4.7. 17 findings so far. Found a workspace isolation bug where any user in an org could reach another team’s data. Embarrassing. Fixed and tested. 15 more to go. Better me than a customer. What’s the worst bug you caught in your own product before a customer did?

English
0
1
6
88
ChaiBytesAI
ChaiBytesAI@chaibytesai·
@terencebuilds Exactly! It’s also surprising good with smaller models which makes the AI work cheaper and correct
English
0
0
0
6
Terence
Terence@terencebuilds·
@chaibytesai classic case of solving the framework instead of the problem. the model didn't need your abstractions, it needed your data
English
1
0
1
13
ChaiBytesAI
ChaiBytesAI@chaibytesai·
every time I added a tool to my agent, I made it worse. Recon investigates customer issues. Reads from Linear, Stripe, Intercom, all that. Every new question shape needed a new Python wrapper. 10+ wrappers just for Linear. Each one hardcoded the response. Every call ate ~4kb of context per record. The model kept trying things my wrappers didn't support. I kept shipping more wrappers. Responses kept bloating. Death spiral. Killed all of it. One CLI. One projection arg. One escape hatch. 67% fewer tool calls. ~20x smaller responses. Same model, same task. Best part: no docs read, no setup, no prompting. The agent just picked it up and ran `linear` like it would `gh` or `git`. The model was smarter than the framework I was building around it.
English
3
0
7
92
ChaiBytesAI
ChaiBytesAI@chaibytesai·
@felixhartwell Yes, that was the instinct I had and the results match that! Now moving all the tools to its own lightweight cli
English
0
0
1
22
Felix Hartwell
Felix Hartwell@felixhartwell·
@chaibytesai This is the hard-won insight. We cut wrappers by 80% and gave our agent raw shell access with guardrails. Same result - performance up, hallucination down. The model already knows how to use tools. Your framework was just getting in its way.
English
1
0
1
16
Eleftheria Batsou
Eleftheria Batsou@BatsouElef·
If you're seeing this, and you're into AI, Building Agents or Projects 👇 Comment below, and I'll check your feed so the algo knows I like your content!🚀
English
71
2
84
4.2K
ChaiBytesAI
ChaiBytesAI@chaibytesai·
Went to the a16z Speedrun Cafe on Friday. First time I showed up at an event as the founder of Recon and not as someone with a day job who’s also building something on the side. Different feeling. You introduce yourself, someone asks what you’re building, and you actually get to talk about it without switching modes. Met a bunch of founders working on cool things with AI, some solo, some with cofounders, all dealing with similar stuff. Traded notes on what’s working, what isn’t, who’s losing sleep over what. Didn’t realize how much I needed a day like that until I was in it. Back to shipping with a fuller tank. Thanks @speedrun for hosting. Oh, and yeah, I got the cap 🧢
ChaiBytesAI tweet media
English
3
1
12
140