技术推文精选 - 2026-05-05

2026年5月5日科技每日简报

Today's top tech conversations are led by @chatgpt21, whose post about 'Jack Clark Co Founder of Anthr...' garnered the highest engagement. Key themes trending across the top stories include companies, video, revenue, margin, anthropic. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. chatgpt21 (Group Score: 355.6 | Individual: 42.2)

Cluster: 13 tweets | Engagement: 557 (Avg: 162) | Type: Tech

Jack Clark Co Founder of Anthropic:

“I'm writing this post because when I look at all the publicly available information I reluctantly come to the view that there's a likely chance (60%+) that no-human-involved AI R&D - an Al system powerful enough that it could plausibly autonomously build its own successor - happens by the end of 2028. This is a big deal. I don't know how to wrap my head around it.”\n\nQT @jackclarkSF: I've spent the past few weeks reading 100s of public data sources about AI development. I now believe that recursive self-improvement has a 60% chance of happening by the end of 2028. In other words, AI systems might soon be capable of building themselves.

See 12 related tweets

@cgtwts: Co-founder of Anthropic:

“I now believe that recursive self-improvement has a 60% chance of happen...

@AISafetyMemes: Anthropic founder says it's almost over.

How will you spend your remaining months? https://t.co/Vwx...

@kimmonismus: Anthropics Jack Clarke now believes that recurse self-improvement has a 60% change of happening by e...
@cgtwts: Anthropic CEO:

“At Anthropic, we essentially have Claude designing the next version of Claude itse...

@dejavucoder: oh husbant... you said ai research would be safe job https://t.co/kgO2hk3pWU\n\nQT @jackclarkSF: I'v...

2. rohanpaul_ai (Group Score: 160.0 | Individual: 62.6)

Cluster: 4 tweets | Engagement: 768 (Avg: 89) | Type: Tech

Bloomberg: OpenAI launches a $ 10Bn joint venture called “The Deployment Company” to help businesses use its AI.

The new company, The Deployment Company, has raised more than $ 4B from 19 investors, including TPG, Brookfield, Advent, Bain, SoftBank, and Dragoneer.

The basic bet is that AI adoption is no longer mainly a model-quality problem, because many companies already want AI but lack the teams, workflows, data access, security rules, and operating discipline to install it safely inside real business processes.

Private equity firms are useful here because they control or advise large webs of companies, and the report says OpenAI’s partners can reach more than 2,000 portfolio companies and clients.

That turns enterprise AI selling from one-company-at-a-time pitching into a routed distribution system, where OpenAI can package software, consulting, deployment playbooks, and sector-specific use cases across finance, healthcare, coding, operations, and support.

The deeper technical point is that LLMs do not create value just by answering prompts, because they need to be connected to company data, permissions, tools, evaluation systems, and human review loops before they can affect revenue or cost.

Anthropic also is building a similar PE-backed route for Claude, which suggests the next AI race may be less about demos and more about who can industrialize deployment fastest.

bloomberg. com/news/articles/2026-05-04/openai-finalizes-10-billion-joint-venture-with-pe-firms-to-deploy-ai\n\nQT @rohanpaul_ai: WSJ: Anthropic is wrapping up a deal to set up a joint venture with Blackstone, Goldman Sachs, and other Wall Street firms, with the goal of selling AI tools to private-equity backed companies, according to people familiar with the matter.

Anthropic, Blackstone, and Hellman & Friedman are expected to put in about $300M each, while Goldman Sachs is expected to invest about$ 150M.

The new company would work like an AI deployment arm, meaning it would not just sell Claude access, but help companies rebuild workflows around LLMs, from customer support and finance to coding, legal review, and internal research.

Private-equity firms are the target because they own many companies, measure every cost tightly, and can force software changes faster than slow public companies.

Anthropic gets distribution, Wall Street gets a stake in the AI services layer, and portfolio companies become a large testing ground for enterprise AI.

The deeper move is that AI labs are no longer only competing on model quality, but on who controls the path from model to business process.

wsj .com/business/deals/anthropic-nears-1-5-billion-joint-venture-with-wall-street-firms-8f5448ee

See 3 related tweets

@cryptopunk7213: anthropic is going after the $300B consulting sector with a new$ 1.5B consulting arm that seeks to p...
@rohanpaul_ai: RT @rohanpaul_ai: Sam Altman is basically saying OpenAI does not want to be just another high-margin...
@rohanpaul_ai: WSJ: Anthropic is wrapping up a deal to set up a joint venture with Blackstone, Goldman Sachs, and o...

3. wallstengine (Group Score: 157.8 | Individual: 24.5)

Cluster: 8 tweets | Engagement: 145 (Avg: 101) | Type: Tech

Anthropic is teaming with Goldman Sachs, Blackstone, Hellman & Friedman and others on a $1.5B venture to deploy Claude across private equity-owned companies. The new firm will embed engineers inside businesses to redesign workflows and speed AI adoption. https://t.co/cefZIfZxUc

See 7 related tweets

@Reuters: Anthropic nears $1.5 billion AI joint venture with Wall Street firms, WSJ reports https://t.co/teJOD...
@Cointelegraph: 🔥 BIG: Anthropic is finalizing a $1.5 billion joint venture with Blackstone, Goldman Sachs, and Hell...
@WSJ: Exclusive: Anthropic is creating a joint venture with Blackstone, Goldman Sachs and a handful of oth...
@CNBC: Anthropic teams with Goldman, Blackstone and others on $1.5 billion AI venture targeting PE-owned fi...
@business: Anthropic is partnering with Blackstone, Hellman & Friedman and Goldman Sachs to form a new comp...

4. StockMKTNewz (Group Score: 145.4 | Individual: 48.1)

Cluster: 4 tweets | Engagement: 2408 (Avg: 333) | Type: Tech

Amazon $AMZN just posted this:

“Amazon is opening its entire logistics network—freight, distribution, fulfillment, and parcel shipping capabilities—to every business, of all types and sizes” https://t.co/Fz9zf9RNB6

See 3 related tweets

@tobi: RT @amazon: TODAY: Amazon is opening its entire logistics network—freight, distribution, fulfillment...
@StockSavvyShay: $AMZN is opening its full logistics network to businesses of all sizes including freight, distributi...
@wallstengine: Amazon $AMZN launched Amazon Supply Chain Services, opening its freight, warehousing, fulfillment, a...

5. wallstengine (Group Score: 140.4 | Individual: 39.2)

Cluster: 4 tweets | Engagement: 199 (Avg: 101) | Type: Tech

$DUOL Q1’26 EARNINGS HIGHLIGHTS

🔹 Revenue: $292.0M (Est.$ 289M) 🟢; +27% YoY 🔹 EPS: $0.89 (Est.$ 0.75) 🟢 🔹 DAUs: 56.5M; +21% YoY 🔹 Paid Subscribers: 12.5M; +21% YoY 🔹 Total Bookings: $308.5M; +14% YoY 🔹 Adj. EBITDA:$ 83.4M; 28.6% margin

Q2 Guide: 🔹 Revenue: $295.5M; +17.1% YoY 🔹 Bookings:$ 283.5M; +5.8% YoY 🔹 Adj. EBITDA: $71.0M 🔹 Adj. EBITDA Margin: 24.0% 🔹 Gross Margin: ~71.0%

FY Guide: 🔹 Revenue: $1,205M; +16.1% YoY 🔹 Bookings:$ 1,280M; +10.5% YoY 🔹 Adj. EBITDA: $310M 🔹 Adj. EBITDA Margin: 25.7% 🔹 Gross Margin: Trending down to ~69.0% by Q4 🔹 Stock-Based Comp: Nearly 15% of revenue 🔹 Effective Tax Provision: ~18-20%

Other Metrics: 🔹 MAUs: 137.8M 🔹 Subscription Bookings: $268.1M; +15% YoY 🔹 Subscription Revenue:$ 250.9M; +31% YoY 🔹 Advertising Revenue: $20.6M; +15% YoY 🔹 Duolingo English Test Revenue:$ 11.3M; -6% YoY 🔹 In-App Purchases Revenue: $8.4M; -11% YoY 🔹 Course Units Published: 20,500 in Q1’26 🔹 Video Call: Average words spoken per user more than doubled over the last year 🔹 Diluted Shares Outstanding: ~49.7M

Financials: 🔹 Gross Margin: 73.0%; +190 bps YoY 🔹 Net Income: $43.5M; 14.9% margin 🔹 Net Cash From Operating Activities:$ 150.8M; 51.6% margin 🔹 FCF: $147.8M; 50.6% margin 🔹 Cash & Cash Equivalents:$ 1.1B

Capital Return: 🔹 Buybacks: ~ $50.6M / ~514,000 shares repurchased through May 1, 2026 🔹 Authorization:$ 400M repurchase program 🔹 Buybacks represent more than 100% of 2025 net dilution from equity awards

Commentary: 🔸 “In Q1, we executed on the strategy we outlined in the last shareholder letter.”

🔸 “It’s still too early to see the impact in our metrics, but our results so far are in line with what we expected.”

🔸 “Speaking practice, historically the biggest gap for learners on Duolingo, is now a core part of the product.”

🔸 “We have conviction this is ultimately what will make us a larger and more durable company.”

See 3 related tweets

@wallstengine: $LSCC Q1’26 EARNINGS HIGHLIGHTS

🔹 Revenue: $170.9M (Est.$ 164.89M) 🟢; +42.2% YoY 🔹 Adj. EPS: $0.41 ...

@wallstengine: $BLZE Q1’26 EARNINGS HIGHLIGHTS

🔹 Revenue: $38.7M (Est.$ 37.79M) 🟢; +12% YoY 🔹 Adj. EPS: $0.04 (Est...

@wallstengine: $EVER Q1’26 EARNINGS HIGHLIGHTS

🔹 Revenue: $190.9M (Est.$ 180.1M) 🟢; +15% YoY 🔹 Net Income: $18.7M ...

6. rohanpaul_ai (Group Score: 131.1 | Individual: 33.5)

Cluster: 4 tweets | Engagement: 47 (Avg: 89) | Type: Tech

A startup in the Top 6 on Artificial Analysis Text-to-Video Leaderboards

Alongside Alibaba, ByteDance, and xAI. Video Rebirth came out of nowhere. The AI video leaderboard has been exclusively trillion-dollar companies.

Today, a startup just broke in. Video Rebirth. Super realistic results.

Their model BACH just hit Top 6 on @ArtificialAnlys.\n\nQT @ArtificialAnlys: Bach-1.0 Preview from Video Rebirth debuts at #6 on the Artificial Analysis Text to Video Leaderboard (No Audio)!

Bach-1.0 Preview is the latest Text to Video model from @video_rebirth, with similar performance to Vidu Q3 Pro, Kling 3.0 Omni 1080p (Pro), and grok-imagine-video.

Bach-1.0 Preview is intended for broad release later in May.

See example generations from Bach-1.0 Preview in the Artificial Analysis Video Arena below 🧵

See 3 related tweets

@Scobleizer: A startup just broke into the Top 6.

Alongside Alibaba, ByteDance, and xAI.

The AI video leaderb...

@AngryTomtweets: A startup nobody's heard of just appeared on the @ArtificialAnlys AI Video leaderboard.

BACH 1.0, d...

@kimmonismus: A little-known startup just landed on the @ArtificialAnlys AI Video leaderboard, now ranked among th...

7. adamlyttleapps (Group Score: 128.6 | Individual: 38.5)

Cluster: 4 tweets | Engagement: 242 (Avg: 79) | Type: Tech

This is the onboarding flow and paywall https://t.co/JhERKeHprD\n\nQT @adamlyttleapps: New onboarding flow is turning 20% of installs into trials with 31% trials converting to paid

Current biggest leak is App Store listing → install ratio from ads got it to 10% and latest update it's dropped to 2% https://t.co/MwVRfmdbM9

See 3 related tweets

@adamlyttleapps: Worst feeling in the world is releasing a version of the app that gets 10.54% app store listing conv...
@adamlyttleapps: Actually, no

My biggest issue is cost per installation

I think I'm incorrectly combining App Store...

@adamlyttleapps: New onboarding flow is turning 20% of installs into trials with 31% trials converting to paid

Curre...

8. LangChain (Group Score: 125.9 | Individual: 42.3)

Cluster: 4 tweets | Engagement: 195 (Avg: 22) | Type: Tech

Start today → https://t.co/qoQex1VAOw https://t.co/yod5lPZhvz\n\nQT @masondrxy: deepagents-cli is quietly becoming the best place to start coding with open weight models.

we've been investing heavily in making it a harness that's truly model-agnostic, without compromising performance!

different models perform best with different harnesses -- prompts, middleware, settings. our recent profiles API (below) lets you bundle all of that per model, so Kimi, Qwen, GLM, etc. can drive the agent loop just as well as the closed frontier.

more info on profiles https://t.co/H61pJj5iGl

other recent wins worth highlighting:

/agents - swap agent profiles mid-session (coding agent/content writer/custom)
/model - fuzzy switcher w/ live status; OpenRouter, LiteLLM, Baseten, hosted Ollama all built-in
headless mode w/ --json + --max-turns for scripting
--acp to run as an ACP server
/skill:name skills
MCP w/ OAuth

full docs and quickstart ⬇️

See 3 related tweets

@hwchase17: open weight models are having a moment

@masondrxy is making deepagents cli a coding harness that wo...

@sydneyrunkle: use this powerful open harness to harness powerful open models!\n\nQT @masondrxy: deepagents-cli is ...
@LangChain_OSS: RT @masondrxy: deepagents-cli is quietly becoming the best place to start coding with open weight mo...

9. wallstengine (Group Score: 125.3 | Individual: 26.7)

Cluster: 6 tweets | Engagement: 67 (Avg: 101) | Type: Tech

NYT: The White House is weighing a plan to vet new AI models before release, a sharp shift from Trump’s earlier hands-off approach, after Anthropic’s Mythos raised alarms over cyberattack risk and pushed officials to seek first access to powerful models. https://t.co/kmIO9uyd7u

See 5 related tweets

@negligible_cap: *WHITE HOUSE EYES VETTING AI MODELS BEFORE RELEASE, NYT SAYS

Prediction markets on best AI model st...

@negligible_cap: *WHITE HOUSE EYES VETTING AI MODELS BEFORE RELEASE, NYT SAYS

White House vetting AI models: https:/...

@Forbes: White House May Review New AI Models Before Public Release, Report Says https://t.co/WXHNAj6zg3 http...
@Reuters: White house considers vetting AI Models before they are released, NYT reports https://t.co/TngauV5QT...
@theobearman: RT @deredleritt3r: New York Times:

Trump administration is considering the introduction of "governm...

10. Conste11ation (Group Score: 116.2 | Individual: 40.7)

Cluster: 4 tweets | Engagement: 126 (Avg: 51) | Type: Tech

The WEF published a white paper on AI in cybersecurity today with KPMG. 84 organizations. 15 industries. They identify governance, audit trails, and traceable evidence as the critical requirements for autonomous AI systems.

@IBM 's ATOM now handles 95% of security investigations autonomously. The report calls it best practice. The audit trail lives in IBM's infrastructure. That's operator-controlled testimony, not independent proof.

$DAG is part of the AI infrastructure story the market is starting to understand.

Constellation spent years building around verified data, machine-to-machine trust, governance, and auditability. 79% of enterprises report AI agent adoption. Only 2% have scaled it.

The gap is exactly the verification problem Constellation was built to solve including six years of R&D through Project IRON SPIDR with the US Department of Defense.

@wef confirmed the problem today.

The EU AI Act Article 12 enforcement starts in August. The audit trail for autonomous AI decisions needs to exist somewhere no single operator controls. Digital Evidence runs on $DAG. That infrastructure is live.

AI is the catalyst. The thesis didn't change. The market caught up to it. Still early.

See 3 related tweets

@Conste11ation: IBM's ATOM system handles 95% of cybersecurity investigations autonomously. The WEF calls it best pr...
@Conste11ation: RT @MarcoSalzmann80: 🧵 AI in cybersecurity is no longer theoretical.

It is already being used in pr...

@Conste11ation: RT @Dagnum_PI: The World Economic Forum and KPMG just published a white paper on AI in cybersecurity...

11. aakashgupta (Group Score: 114.4 | Individual: 44.8)

Cluster: 3 tweets | Engagement: 226 (Avg: 63) | Type: Tech

$1.4M total comp at Google. He walked away from it.

Mahesh spent 13 years at Microsoft, Amazon, Meta, Google. Doubled his comp every two years through the AI wave. His friends from Meta moved to Nvidia and are at $2.5M now. The recruiters keep calling. The next 70% bump is sitting there.

He left anyway. The reason matters more than the money.

Every meaningful AI product of the last three years was built by a small team or one person. ChatGPT came out of a small company. Lovable started tiny. Claude Code was a small team inside Anthropic. OpenClaw is one developer in Australia.

Inside the $2T company, the two-page document needs a one-page approval that takes six weeks. In those same six weeks, somebody with no team and no permission became a builder PM who can ship anything.

The safe job stopped being safe the moment the tools collapsed the cost of building a product from a six-person engineering team to one person at a laptop.

The $1.4M is the price of staying static while the people around you compound.\n\nQT @aakashgupta: This guy literally broke down how to become a$ 1.4M "builder PM" with n8n, Claude Code, and OpenClaw:

1:53 - What a "builder PM" actually is 6:04 - Your first agent in n8n (live build) 14:18 - Why every agent needs these 4 things 21:35 - The multi-agent eval loop 29:47 - Where n8n dies 33:39 - When to graduate to Claude Code 35:08 - What broke in December 2025 47:17 - The self-improving PRD reviewer 1:02:28 - Mocks and prototypes without designers 1:05:15 - OpenClaw and the new agent OS 1:22:06 - What AI PM interviews look like now

See 2 related tweets

@gregisenberg: Andrew Wilkinson owns 40+ businesses.

He just showed me how he's using OpenClaw, Claude Code and AI...

@VaibhavSisinty: Wild week in AI. I had to triple-check half of these were real. 😨

→ China blocked Meta's $17,000 cr...

12. chhddavid (Group Score: 109.0 | Individual: 27.6)

Cluster: 6 tweets | Engagement: 11 (Avg: 275) | Type: Tech

so you're telling me Claude Opus 4.7 can now...

build a video game
launch it on Steam, Itch
do email marketing
self-build features

without any human in the loop??

it's so over... https://t.co/i5Fe0OlsL5\n\nQT @shipper_now: 🚨 JUST IN: video game vibe coding is here!

build games just by talking to AI https://t.co/wCaDdCnPs9

See 5 related tweets

@chddaniel: oh my... So you're telling me CC w/ Opus 4.7 can now:
clone any game
publish on Steam, Itch
d...
@chddaniel: game developers right now: https://t.co/VNPXjm43J0\n\nQT @shipper_now: 🚨 JUST IN: video game vibe co...
@chddaniel: babe wake up.

prompt to $50M video game is literally one click away now. https://t.co/NIKK9bgkno\n\...

@chddaniel: this is f*king scary.....\n\nQT @shipper_now: 🚨 JUST IN: video game vibe coding is here!

build game...

@chhddavid: this is terrifying.\n\nQT @shipper_now: 🚨 JUST IN: video game vibe coding is here!

build games just...

13. nvidia (Group Score: 101.5 | Individual: 34.3)

Cluster: 3 tweets | Engagement: 74 (Avg: 394) | Type: Tech

AI isn't just software. It's a five-layer cake: energy, chips, infrastructure, models, and applications. The countries and companies that build the full stack will define the next industrial era. Read more here: https://t.co/Lc9Cfjg2nv\n\nQT @TheFuturumGroup: AI investment is expanding beyond models and applications into the full stack that supports them. Energy capacity, semiconductor supply, compute infrastructure, and software layers are becoming more connected, influencing how and where AI can be built and deployed.

In partnership with @nvidia, @TheFuturumGroup examines the Five Layers of the AI Cake framework as a way to understand the structure of the AI economy. The market brief outlines how value is distributed across the stack, where bottlenecks are forming, and what these dynamics mean for enterprise strategy, capital allocation, workforce planning, and global competitiveness.

⏯️ Explore the key findings, and use the link to download the full report:
👉 https://t.co/9gtyfE2w6U

#AIInfrastructure #NVIDIA #Futurum

See 2 related tweets

@nvidianewsroom: AI isn't just software. It's a five-layer cake: energy, chips, infrastructure, models, and applicati...
@StockSavvyShay: TOP 10 AI REVENUE WINNERS

AI revenue is concentrating around the companies that either build the co...

14. Scobleizer (Group Score: 99.0 | Individual: 62.7)

Cluster: 2 tweets | Engagement: 2874 (Avg: 115) | Type: Tech

RT @Malay4Product: Let me explain what just happened today because it deserves so much recognition.

GalaxEye is a Bengaluru startup founded in 2021 by IIT Madras engineers. Today they launched Mission Drishti on a SpaceX Falcon 9. It is India's largest privately built satellite at 190 kg. And it carries a technology that no commercial satellite has ever carried before.

Normal satellites take photos of the Earth using optical cameras. Like your phone camera, but from 500 km up. The problem is obvious. Clouds. Night. Fog. Smoke. If any of these are in the way, the photo is useless. India has monsoon cover for 4 months a year. That is 4 months where optical satellites are partially or fully blind over large parts of the country.

The alternative is SAR. Synthetic Aperture Radar. Instead of taking photos with light, it sends radar waves down and reads what bounces back. Radar goes through clouds, through darkness, through smoke. A SAR satellite can image a flooded village at 2 AM during a cyclone when no optical satellite can see anything.

The problem with SAR is that the images look nothing like photos. They look like grainy black-and-white radar maps. A military analyst or a trained geospatial engineer can read them. A farmer, a disaster response team, or a city planner cannot.

Until today, if you wanted both optical and SAR data for the same location, you needed two different satellites, passing over at different times, at different angles. Then someone had to manually align and fuse the two datasets. Expensive, slow, and the data never perfectly matched because the satellites saw the same spot minutes or hours apart.

GalaxEye put both sensors on one satellite. Optical and SAR, fused into what they call OptoSAR. Three times more information than a single sensor. Processed onboard by an NVIDIA AI chip at 1.8 metre resolution.

Now in practice, during the next cyclone hitting Odisha, one satellite pass gives you a clear image of which villages are flooded, which roads are cut, and which buildings are standing. Day or night. Cloud or clear. In near real-time.

For defence, it means you can monitor a border area 24/7 regardless of weather. For agriculture, it means tracking crop health across an entire monsoon season without a single cloud gap. For infrastructure, it means monitoring construction progress on highways and bridges without waiting for a clear day.

GalaxEye tested their SAR tech on ISRO's POEM orbital platform. The satellite was tested at ISRO facilities. IN-SPACe provided regulatory clearance. NSIL, ISRO's commercial arm, will distribute the imagery globally. And it launched on SpaceX because ISRO's PSLV doesn't have the right orbit slot for this mission.

Yes, four IIT Madras graduates built a world-first satellite in 4 years in Bengaluru.

Take a bow!

See 1 related tweets

@chandrarsrikant: After Drishti launch,GalaxEye eyes plans 10 more, new facility

At roughly 190 kg, Drishti is among ...

15. FirstSquawk (Group Score: 98.9 | Individual: 28.4)

Cluster: 4 tweets | Engagement: 182 (Avg: 83) | Type: Tech

OPENAI COMPLETES $10 BILLION PARTNERSHIP WITH PRIVATE EQUITY FIRMS TO LAUNCH AI TECHNOLOGIES.

See 3 related tweets

@business: OpenAI has raised more than $4 billion for a new joint venture that will focus on helping businesses...
@Techmeme: Source: OpenAI has raised over $4B at a$ 10B pre-money valuation for The Deployment Company, a new j...
@FirstSquawk: OPENAI HAS SECURED OVER $4 BILLION FOR A NEW JOINT VENTURE TO ASSIST BUSINESSES IN IMPLEMENTING ITS ...

16. teortaxesTex (Group Score: 92.5 | Individual: 30.2)

Cluster: 5 tweets | Engagement: 84 (Avg: 75) | Type: Tech

Clear article on bipartisan consensus, with some mandatory chyna hypocrisy. We all know American-Israeli AI ambitions, luridly advertised at the start of this war in Iran, for example. I'll be plain, these people shouldn't «control A.I.’s future» if humanity is to ever be free. https://t.co/yO4HYmwvrl\n\nQT @deanwball: Today, @BuchananBen and I co-author a piece in the New York Times with a simple message:

While we disagree on plenty, we believe AI has national security implications which deserve a careful and bipartisan government response. We can (and should) have partisan fights about all manner of AI issues, but catastrophic risk from AI shouldn’t be one of them.

See 4 related tweets

@AlecStapp: Important piece in the NYT today from @deanwball & @BuchananBen.

Mythos found zero-day vulnerab...

@AndrewCurran_: Dean Ball this morning in the NYT. https://t.co/AMRlJPtAGi\n\nQT @deanwball: Today, @BuchananBen and...
@Miles_Brundage: Hello, based police\n\nQT @deanwball: Today, @BuchananBen and I co-author a piece in the New York Ti...
@Miles_Brundage: RT @deanwball: Today, @BuchananBen and I co-author a piece in the New York Times with a simple messa...

17. StockMKTNewz (Group Score: 90.5 | Individual: 25.3)

Cluster: 4 tweets | Engagement: 283 (Avg: 333) | Type: Tech

Cerebras just gave out some more details for its upcoming IPO

Cerebra said today its planning to raise ~ $3.24 billion at a post IPO valuation of ~$ 25.5 billion\n\nQT @IPONewsroom_: JUST IN: CEREBRAS SYSTEMS, THE AI CHIP COMPANY THAT BUILDS THE WORLD'S LARGEST COMMERCIAL PROCESSOR, JUST FILED AMENDMENT NO. 1 TO ITS S-1 WITH FULL PRICING TERMS

The deal: • 28,000,000 shares offered at $115-$ 125 per share • Estimated raise: $3.24 billion at the midpoint, or$ 3.73 billion if underwriters exercise the over-allotment option in full • Estimated post-IPO valuation: ~ $25.5 billion at the midpoint • Listing on the Nasdaq Global Select Market under ticker$ CBRS • 4,200,000 share over-allotment option

The financials: • Revenue: $510M in 2025, up 76% from$ 290M in 2024 • 10x revenue growth over 3 years (from $24.6M in 2022) • First GAAP profitable year:$ 237.8M net income in 2025 • Non-GAAP net loss of $75.7M in 2025

The marquee customer: • Multi-year deal with OpenAI valued at more than $20 billion, announced January 2026 • OpenAI agreed to deploy 750 megawatts of Cerebras compute • OpenAI provided Cerebras with a$ 1 billion Working Capital Loan • AWS signed a binding term sheet in March 2026 to become the first hyperscaler to deploy Cerebras

The technology: • Cerebras Wafer-Scale Engine 3 (WSE-3) is 58 times larger than NVIDIA's B200 chip • 4 trillion transistors, 900,000 compute cores, 44 gigabytes of on-chip memory • Up to 15x faster than GPU-based solutions on inference benchmarks

Lead bookrunners: Morgan Stanley, Citigroup, Barclays, UBS Investment Bank.

Co-managers: Mizuho, TD Cowen, Needham & Company, Craig-Hallum, Wedbush Securities, Rosenblatt, Academy Securities, Credit Agricole CIB, MUFG, First Citizens Capital Securities.

Cerebras was founded in 2016 by Andrew Feldman, who serves as CEO and President.

See 3 related tweets

@Reuters: Cerebras targets $26.6 billion valuation in US IPO as AI chip demand surges https://t.co/2WGCj248sX ...
@StockSavvyShay: Cerebras $CBRS filed updated IPO terms offering 28M shares at$ 115- $125 to raise ~$ 3B at ~$26B valua...
@business: Cerebras is seeking to raise as much as $3.5 billion in an IPO, as the artificial intelligence chipm...

18. aakashgupta (Group Score: 86.4 | Individual: 35.0)

Cluster: 3 tweets | Engagement: 5 (Avg: 63) | Type: Tech

AI product sense is now a real interview round. And the candidates treating it like "regular product sense with AI buzzwords sprinkled in" are losing offers right now.

Three tiers of companies are running this round. Each tier wants something different. Walking in with the wrong prep is how strong PMs get blindsided.

Tier 1: AI-native companies.

OpenAI, Anthropic, Google DeepMind. Dedicated AI product sense round in every PM loop. Not optional. Not team-specific. Every candidate gets it.

The questions force model-aware tradeoffs. Ankit shared an OpenAI example: "How would you double ChatGPT image creation with just three engineers?" You're allocating against model constraints, engineering scarcity, and product bets at the same time. Standard product sense playbooks weren't built for this.

Tier 2: Big tech that just bolted it onto the loop.

Meta is the move to watch. Ankit shared that they added a fourth interview specifically for this. Live prototyping. Interviewer in the room. Meta calls it "vibe coding" internally. You ship while they watch.

Google's AI teams, Amazon's https://t.co/gPKaDZV38V org, and NVIDIA are running versions of this. Often gated to specific teams or seniority levels, but the format is the same: prototype live, defend choices in real time.

Tier 3: Companies weaving AI into the existing product sense round. This is the trap.

LinkedIn won't label it as an "AI round." But the questions in their standard product sense slot now sound like: "How would you leverage https://t.co/gPKaDZV38V capabilities while thinking through this product?" Or: "What are the implications of OpenAI and Anthropic capturing new markets, and how should our product rethink its strategy?"

Answer those without referencing model capabilities or competitive market shifts and you fail. Quietly. You'll walk out thinking you nailed the case. The rejection email won't tell you what happened.

The skill split is real:

Tier 1 wants you to reason like a model-aware operator under engineering constraints.

Tier 2 wants you to ship working prototypes live with AI tools.

Tier 3 wants AI reasoning baked into every product answer without being prompted.

Most candidates I've seen are still prepping for the 2022 version of product sense. The ones converting offers at companies like Uber, Atlassian, and Cisco are training on what the loop actually looks like now.

The full mock with Ankit shows what a tier 2 vibe coding round looks like end-to-end, plus the 9/10 breakdown of where it landed.\n\nQT @aakashgupta: AI PM interviews are now testing "AI product sense."

So I recorded a mock to demystify what it is with Ankit Virmani, who just nabbed AI PM offers at Uber, Atlassian, and Cisco.

6:59 3 tiers running it 12:04 Live mock 52:30 9/10 breakdown https://t.co/Oinn6sNmok

See 2 related tweets

@aakashgupta: I've been telling PM candidates this for months: AI product sense is no longer a niche round.

12 mo...

@aakashgupta: The behavioral round gets you through the door. AI product sense decides the size of your offer.

Mo...

19. BrianRoemmele (Group Score: 85.9 | Individual: 35.7)

Cluster: 3 tweets | Engagement: 297 (Avg: 314) | Type: Tech

How I Built A Robot Vision System in the 1980.

I built this robot in 1984. It uses a Polaroid Ultrasonic Ranging System Designer’s Kit for distance measurements. This technology absolutely fascinated me when I first saw it on the camera and it started me on a few businesses in 1980 buying 1000s of these sensors from the commercial division of Polaroid for two product I built.

One was a radio and battery based truck backup sensor that I sold via a relationship I made with two insurance companies.

See the damage done by trucks in the 1980s to shipping loading docks were a big thing millions of dollars in insurance claims, and the companies would bill the trucking company insurance $1000 or more per impact.

So I knew this and called two of the largest commercial insurance companies in the US and hacked my way to “the person” in one and than used that to get to the other (you competitor just talked to me…).

“Hey I am in my garage and I am grifting for like and subscribes this idea”

I mean “Hello, you paid $743,281 last quarter on loading dock damages. How would you like that to be$ 0?” Ha, do I have your attention now! No one ever hung up the phone after that.

My product was “SafeReverse” and I sold 1000s at ~$1,400 per from my garage. I later made a version for front and side, but had less takers. The insurance companies required it for some accounts and lowered premiums for others if they installed it. They made all the marketing materials. I made the product and money.

I hand made everyone of them and hired no one.

I used a 6502 processor board, a radio data module and a SST text to voice synthesizer chip. It would say “10 feet, 2 inches… 5 feet, 6 inches, 11 inches, speed check warning, and “slow down” at 9 inches if the rate of acceleration was too high. I had a printer that would print a receipt incase there was any claims.

It took me 12 minutes to build them from my parts.

I am told I saved over $12 million dollars in claims at one insurance company.

I sold the company in 16 months and found out the acquirer had no intention of making these. I still don’t know why but have suspicions.

Polaroid was so supportive of my efforts that they offered testing equipment and chip sets to make it easier. They never made me feel like they would crush me. They never care “who I was”. I had direct engineering phone numbers.

The company flew me out after the acquisition and wanted to hire me to find more niches for the system. Unfortunately they were already cutting back on to many levels and I saw the reality that was coming.

So this is my history in robotic “vision” before anyone understood what it was.

I have an SX70 kit in a picture frame to remember.

It is in my garage where I still am, typing this to you…

See 2 related tweets

@BrianRoemmele: I purchased my first my Polaroid Ultrasonic Ranging System for $99 (student discount) in 1980.

The ...

@BrianRoemmele: “Shall we play a game?” https://t.co/ZeQfaGh6lz\n\nQT @BrianRoemmele: I purchased my first my Polaro...

20. Shipper_now (Group Score: 85.1 | Individual: 29.7)

Cluster: 3 tweets | Engagement: 28 (Avg: 25) | Type: Tech

this Claude thing actually got scary...\n\nQT @chhddavid: Introducing Website to App. Turn any website into a native mobile app.

Just paste the URL.

Claude Code Opus 4.7 will build, design, launch and translate a mobile app inspired by the original website.

We’ve been using this internally a ton for iOS/Android apps. https://t.co/uwIvNCJAeQ

See 2 related tweets

@chddaniel: this is genuinely terrifying...\n\nQT @chhddavid: Introducing Website to App. Turn any website into ...
@RoundtableSpace: YOU CAN NOW DESIGN, BUILD, AND LAUNCH A NATIVE MOBILE APP ALL FROM PASTING A WEBSITE URL

https://t....