X / Twitter
Replit CEO Amjad Masad
Replit CEO Amjad Masad said developers building on Replit have a clear competitive advantage from the speed of parallel agents, bluntly calling anyone competing with them disadvantaged. He also humorously contrasted this with complicated scraping schemes for obtaining free ChatGPT subscriptions, suggesting simpler approaches are often more efficient.
https://x.com/amasad/status/2033007930669187333
https://x.com/amasad/status/2033103499840434607
Y Combinator CEO Garry Tan
Y Combinator CEO Garry Tan described a breakthrough from using the CLI’s /qa and /browse features in gstack: fully escaping a black-box testing process. He argued that AI is creating an astonishing future, though only a minority truly believes it yet. He also observed that once an implementation in one codebase is understood, reproducing its logic elsewhere takes only around half an hour, making code hoarding outdated. He continued calling for San Francisco to start building again, while gstack reached number 8 on a leaderboard.
https://x.com/garrytan/status/2033307654660018240
https://x.com/garrytan/status/2033316355743576566
https://x.com/garrytan/status/2033322040569065524
https://x.com/garrytan/status/2033326553812074582
https://x.com/garrytan/status/2033326832322334970
Every CEO Dan Shipper
Every CEO Dan Shipper emphasized a core principle for AI-era product building: models improve so quickly that teams must be willing to start over every few months. He sees this as an advantage for small AI teams, whose restart costs are much lower. He also described his Codex workflow: one main conversation handles all the day’s tasks, with pins tracking work in progress not yet deployed. He said coding agents’ GUI/UX urgently needs redesigning.
https://x.com/danshipper/status/2033198890338398318
https://x.com/danshipper/status/2033201069279310330
https://x.com/danshipper/status/2033213845192294793
FPV Ventures Partner Nikunj Kothari
FPV Ventures Partner Nikunj Kothari revealed that 4 companies he backed, 3 angel investments and 1 fund investment, had been acquired in recent months. He also sees more founders seeking M&A advice, including how to assess terms for joining a large company, and offered help to US founders outside crypto and defense.
https://x.com/nikunj/status/2033186390238998690
OpenClaw founder Peter Steinberger
Peter Steinberger, the self-described ClawFather, shared several OpenClaw developments: the next version will let users ask agents questions while they are busy; he is considering a more powerful plugin system with a slimmer core, with Claude Code and Codex plugin-bundle support planned; an AI-built scheduled task runs every 5 minutes to clean up Twitter mentions and filter spam replies, working better than expected; and an MCP-supporting branch is in testing.
https://x.com/steipete/status/2033050183492382767
https://x.com/steipete/status/2033215216469614923
https://x.com/steipete/status/2033220404555526154
https://x.com/steipete/status/2033231158146744582
Peter Yang (AI tutorial creator)
Peter Yang shared his podcast with Ramp CPO Geoff, centered on “my job is to automate my job,” which he sees as everyone’s mission from now on. He called Ramp the most complete example of an AI-native company he has seen. He also shared two scenarios AI produced when asked about white-collar employment, with the optimistic forecast seeing recovery in 2035, and reflected on technical careers: the best direction is not promotion into management but becoming the best builder in the world.
https://x.com/petergyang/status/2033189089315635310
https://x.com/petergyang/status/2033207549785477205
https://x.com/petergyang/status/2033241918562513303
Matt Turck (FirstMark Capital VC)
FirstMark Capital VC Matt Turck captured the compression of product planning in a short post: roadmaps spanned 1 year in 2022, 6 months in 2024, and 1 week in 2026. He added that by 2028, agents might even decide the direction themselves.
https://x.com/mattturck/status/2033293364360052946
https://x.com/mattturck/status/2033295813074690240
Builder Zara Zhang
Zara Zhang observed a new trend: rather than subscribing separately to various model-wrapper products, more people buy model API subscription access directly and obtain personalized experiences through agent skills. She also commented wistfully on Qwen’s former place in the S-tier of models.
https://x.com/zarazhangrui/status/2033039119287455862
https://x.com/zarazhangrui/status/2033076228044067197
South Park Commons General Partner Aditya Agarwal
South Park Commons General Partner and former Dropbox CTO Aditya Agarwal said the most striking thing about knowledge workers who use AI deeply is that they enjoy their work. This is unsurprising: when people remain in flow and feel the satisfaction of solving interesting problems, work itself becomes enjoyable.
https://x.com/adityaag/status/2033184840435314858
swyx (AI developer)
swyx expressed shock at Qwen’s direction. As a former S-tier AI model family, its demise was hard for him to process. He also noted Bolt’s ultimate solution: rather than rent cloud compute, use the more powerful local computer already sitting on every desk.
https://x.com/swyx/status/2033030744352993296
https://x.com/swyx/status/2033228801103298821
Podcasts
Lenny's Podcast — The tactical playbook for getting 20-40% more comp (without sounding greedy) | Jacob Warwick
Guest Jacob Warwick is a professional negotiation adviser who works behind the scenes for Fortune 500 executives, professional athletes, and Hollywood celebrities, helping clients secure more than $1 billion in additional compensation. He rarely appears publicly, making this a rare in-depth interview.
Negotiation starts much earlier than you think
Jacob emphasized that negotiation begins with the first line of your LinkedIn profile. Photo quality and the amount of public information already put a price tag on you. He has repeatedly estimated someone’s salary accurately from their LinkedIn photo alone. His conclusion: the less public information available, the more control you have over the narrative on a call.
The most common expensive mistake: negotiating by email
Many people make requests by email because negotiation feels uncomfortable, but email offers no control over tone. A CEO reading it in a bad mood at airport security might simply think, “This person wants more money again.” Jacob recommends video at minimum, ideally meeting in person or inviting the other party for a walk. Walking side by side feels psychologically collaborative rather than confrontational and can improve results.
“Could there be a little more?” is the simplest, most effective first step
Jacob said that even without a professional negotiator, asking “Is there any possibility of improving on this?” almost always produces a 20% improvement. When he intervenes, his average target is 40%, with the largest cases reaching 300% to 400% of the original offer. His view is that companies usually extract 5 to 100 times your salary in value; asking for more is fair, not greedy.
Do not anchor a number too early
When recruiters ask for salary expectations, Jacob recommends responding, “What is the range for this role?” He moved from $14 an hour to $120,000 a year when younger using that question: do not reveal your floor, let them offer first, then affirm that “I am at the high end.” Once stated, your anchor becomes a ceiling; if the other side speaks first, you never know how high it might go.
“Sell the vacation”: let the other side picture a painless future
Jacob shared a core framework: in an interview, do not start by describing your past. Ask questions, identify the other party’s pain, then guide them through a picture: “Imagine we have worked together for six months and you stand tall while reporting to the board. How does that feel?” Once they connect you with a painless, successful future in their imagination, competitors disappear.
Slowing down increases your power
Manufactured urgency, such as “we need your answer this week,” is your enemy. Jacob recommends deliberate delays: wait a few days to answer emails, control meeting windows by negotiating only when most alert, and even request moving a meeting from the other party’s office to neutral ground. “When you have the power to say no, you have unlimited negotiating power.”
Performance triggers are a powerful way past salary ceilings
When base salary hits the top of the band, Jacob recommends milestone-based triggers, similar to Tom Brady’s Tampa Bay contract: $500,000 for 8 wins, $1.5 million for 12, and more for reaching the playoffs. Such clauses are easier for employers to accept psychologically and identify people genuinely willing to take ownership.
How to get immediate feedback and advance the next conversation
Jacob recommends asking directly before an interview ends: “Is this a done deal? Is there any area where other candidates are stronger than me?” Get feedback immediately rather than follow up after a rejection. If concerns remain, schedule the next conversation while things are fresh instead of surrendering initiative.