- Published on
Top Tech Tweets - 2026-06-20
- Authors
- Name
- geeknotes
Tech Daily Briefing for 2026-06-20
Today's top tech conversations are led by @MTSlive, whose post about 'SITUATION DETECTED: John Jumpe...' garnered the highest engagement. Key themes trending across the top stories include models, briefcase, china, model, claude. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.
1. MTSlive (Group Score: 341.8 | Individual: 48.5)
Cluster: 15 tweets | Engagement: 1343 (Avg: 120) | Type: Tech
SITUATION DETECTED: John Jumper, who won the Nobel Prize in Chemistry for AlphaFold along with Demis Hassabis, has left Google DeepMind to join Anthropic.
See 14 related tweets
- @rohanpaul_ai: John Jumper, the AlphaFold scientist who shared a Nobel Prize with Demis Hassabis, is leaving Google...
- @business: Google DeepMind Vice President John Jumper, who won the 2024 Nobel Prize in chemistry for his work o...
- @kimmonismus: What the fu*k!! John Jumper, one of the key figures behind AlphaFold and co-winner of the Nobel Priz...
- @ns123abc: “Sir… John Jumper… the director of DeepMind… the co-creator of AlphaFold… the man who won the Nobel ...
- @Yuchenj_UW: John Jumper is a Nobel laureate who won the 2024 Nobel Prize with Demis Hassabis.
Now he’s leaving ...
2. swyx (Group Score: 255.2 | Individual: 34.4)
Cluster: 9 tweets | Engagement: 127 (Avg: 122) | Type: Tech
Anthropic is going to IPO at $2T\n\nQT @JohnJumperSci: A bit of news: After nearly 9 years, I have decided to leave Google DeepMind and join Anthropic (after taking some time to recharge). I am incredibly grateful for my time at GDM. @demishassabis took a real chance letting me lead the AlphaFold team just six months after finishing my PhD, and the entire GDM team taught me so much about how to do great science. GDM is a special place, and I’ll still be excited to hear about what amazing things they discover next.
See 8 related tweets
- @zephyr_z9: Anthropic is beefing up its Biology/Chemistry AI team\n\nQT @JohnJumperSci: A bit of news: After nea...
- @demishassabis: Thanks John for an extraordinary partnership and wonderful collaboration over the past 9 years! What...
- @dejavucoder: excited for anthropic biology arc\n\nQT @JohnJumperSci: A bit of news: After nearly 9 years, I have ...
- @dejavucoder: not sure if TBPN can post this now so let me post on their behalf https://t.co/UfXWLlbvwG\n\nQT @Joh...
- @Vtrivedy10: pretty incredible news, for some inspiration of how impact John’s work is, highly recommend watching...
3. kimmonismus (Group Score: 190.9 | Individual: 35.9)
Cluster: 8 tweets | Engagement: 507 (Avg: 510) | Type: Tech
I think many people still underestimate the full implications of China potentially gaining access to ASML-level EUV technology.
The supply chain currently works like this:
-ZEISS in Germany produces the ultra-precise optics and mirrors that are essential for ASML’s lithography systems.
-ASML is the only company in the world that can supply EUV lithography machines at industrial scale. These machines use extreme ultraviolet light to expose tiny chip structures onto silicon wafers, enabling the production of the most advanced semiconductors.
-The machines are shipped in many separate modules, then assembled and calibrated at fabs such as TSMC in Taiwan. This is where companies like NVIDIA, AMD and Apple have many of their most advanced chips manufactured.
The key point is this:
EUV is not just one machine. It is an entire industrial ecosystem. Optics, light sources, metrology, software, precision engineering, service teams and years of manufacturing know-how all have to work together. That is why there is no real alternative to ASML for leading-edge chip production at scale.
This is also why China has been denied access to EUV for years.
China has made enormous progress in AI chips and semiconductor manufacturing, especially through Huawei and SMIC. But it still lacks the mature industrial ecosystem needed to replicate ASML’s EUV capabilities at scale. Huawei’s Ascend chips are improving fast and are now estimated to reach roughly 60 to 80 percent of H100-class performance in some workloads, but they still lag behind NVIDIA in software, memory, reliability and large-scale training infrastructure.
That is why the latest reports are so important.
Reuters previously reported that a Chinese team, including former ASML engineers, had built an EUV prototype in Shenzhen. Now, the US government is reportedly concerned that China may have gained access to ASML’s most sensitive lithography technology. ASML denies that it has ever shipped EUV machines or EUV-specific components to China.
If China really has gained meaningful access to EUV know-how, this would put the West under serious pressure.
It would not mean China can suddenly mass-produce cutting-edge chips tomorrow. But it could shorten the timeline, accelerate Huawei’s chip roadmap and weaken one of the West’s most important technological chokepoints.
Many people are missing the significance of this.
EUV is not just about chipmaking. It is one of the deepest bottlenecks in the global AI race.
(Image: ASML EUV)\n\nQT @kimmonismus: ASML’s EUV machines may have made their way into China, prompting fresh scrutiny from the Trump administration over possible violations of U.S.-led export restrictions.
A similar alarm was already raised in late 2025, when Reuters reported that former ASML engineers had helped China build an EUV prototype in Shenzhen by reverse-engineering ASML’s technology.
This is no small matter. If China finds a way to copy ASML's technology, they could rapidly catch up in chip production and reach parity with the top Western companies.
See 7 related tweets
- @rohanpaul_ai: ASML just became the center of a US-China chip fight after Washington said it fears a banned EUV lit...
- @bubbleboi: This is all lies.
It’s a Frankenstein machine using a canon stage and an old EUV source they stole...
- @theobearman: USG thinks an ASML EUV machine is in China. This is incredibly concerning news if true, with any suc...
- @MarioNawfal: 🇺🇸🇨🇳🇳🇱 U.S. Commerce Sec. Lutnick has warned chip maker ASML that Washington is concerned one of its...
- @StockSavvyShay: $ASML says U.S. officials raised concerns that one of its advanced EUV lithography systems may have ...
4. benitoz (Group Score: 182.9 | Individual: 33.3)
Cluster: 6 tweets | Engagement: 87 (Avg: 48) | Type: Tech
People don’t leave because the lab got dumb.
They leave because the compute got political.
Google has DeepMind fighting internal allocation battles while Anthropic gets capacity. That’s the trillion-dollar irony.\n\nQT @synthwavedd: 🚨 SCOOP: After the release of Fable 5 and with GPT-5.6 looming, the mood behind the scenes at Google DeepMind is increasingly one of frustration and broad discontent over the lab's perceived fall into a distant third—or even fourth—place.
"I can't blame Noam [Shazeer] for walking. He won't be the last big name to go, either," a well-connected DeepMind employee told me.
DeepMind's last major model release, 3.5 Flash, was a significant jump over its predecessor; however, it was not meaningfully better in most cases than 3.1 Pro, released back in February. In real-world use, it remains several steps behind the frontier. That was four months ago, and Google's best model now sits in a lowly fifth place on the Artificial Analysis Intelligence Index—lapped by models from Anthropic, OpenAI, and now China's Zhipu AI. Other releases have proven similarly disheartening: the small video generation model Gemini Omni Flash launched to little fanfare and was easily beaten by ByteDance's Seedance 2.
Gemini 3.5 Pro, slated to launch June 30th, is "not the step change we need to be truly competitive in the race [to AGI]," per another individual at the company. The consensus seems to be that leadership at Google has all but conceded that race to Anthropic and OpenAI, and that "only a big shake-up" will propel them back to the highs of mid-to-late 2025.
But employees are not hopeful: "We no longer have a frontier model in text, image, video, voice, or even vision... if we can't release a real frontier model after over four months of work with all of these resources, what are we doing?"
See 5 related tweets
- @zephyr_z9: Things are not looking good for Google They have missed the agentic coding train Multimodal/always o...
- @MTSlive: SITUATION ANALYSIS: Is OpenAI Back?
This week, star researcher @NoamShazeer announced that he’s joi...
- @teortaxesTex: It was both after all I can't find the original post about Google Anyway, somewhere in the late 2023...
- @Hesamation: 🚨 Google DeepMind is struggling to compete with Anthropic and OpenAI as Gemini 3.5 Pro doesn’t seem ...
- @inafried: RT @synthwavedd: 🚨 SCOOP: After the release of Fable 5 and with GPT-5.6 looming, the mood behind the...
5. teortaxesTex (Group Score: 167.6 | Individual: 37.5)
Cluster: 5 tweets | Engagement: 119 (Avg: 78) | Type: Tech
Zhipu post-training team is very very good\n\nQT @ArtificialAnlys: Announcing AA-Briefcase, the benchmark for the next era of agentic knowledge work
AA-Briefcase is our new benchmark for testing models on long-horizon knowledge work tasks in complex projects built by industry experts. Models are evaluated on multi-week projects, each with many linked tasks and thousands of input source files.
We evaluated Claude Fable 5 from @AnthropicAI before it became unavailable, and it currently leads with an Elo score of 1587, followed by Claude Opus 4.8 (max, 1356), Opus 4.7, and the recently-released GLM 5.2 (max, 1266) from @Zai_org.
Claude Fable 5 cost 10.40, GPT-5.5 (xhigh) at 2.40.
AA-Briefcase comprises four private scenarios, each representing a multi-week knowledge work project set in a realistic organizational context. A public fifth scenario has been released via @huggingface as a representation of scenario structure, submission, and grading (AA-Briefcase Lite). This does not count toward official AA-Briefcase results, and is demonstrative only.
Key elements of AA-Briefcase:
➤ Realistic long-horizon projects: AA-Briefcase moves beyond single, disconnected prompts by evaluating models across a coherent long-horizon project. Tasks build week by week, draw on shared institutional context, and require deliverables such as financial models, board presentations, and design mock-ups
➤ Large volumes of fragmented context: AA-Briefcase requires models to reason across thousands of inputs, including company documents, meeting transcripts, large-scale data exports, 25,000+ Slack messages and 3,500+ emails. These sources are fragmented, messy, and often contain realistic contradiction, testing whether models can navigate the ambiguity of real-world knowledge work
➤ Composite rubric and pairwise grading: AA-Briefcase combines binary rubric checks for ground-truth correctness with pairwise grading on analytical quality and presentation quality. Unlike many evaluations that focus on a single metric, AA-Briefcase tests agentic capabilities more comprehensively, exposing cases where models produce outputs that look polished but are incorrect or lack analytical rigor
➤ Built by industry experts: AA-Briefcase scenarios mirror real-world knowledge work, with tasks developed over months by experts across data science, product management and corporate strategy from companies including Google, McKinsey & Company and BCG. Task challenges are drawn from professional experience, making AA-Briefcase more reflective of the ambiguity, messy context and competing priorities that define real-world knowledge work
Key results:
➤ Claude Fable 5 leads AA-Briefcase at 1587 Elo: This is followed by Claude Opus 4.8 (1356) with the next-best non-Anthropic model, GLM-5.2 (max), ~90 points back at 1266. Note that Claude Fable 5 did not use the Opus 4.8 fallback for any task in AA-Briefcase
➤ Cost per task varies by ~800x across models tested: Claude Fable 5 leads the benchmark but costs more than 0.04 for DeepSeek V4 Flash (max). The strongest price/performance options are open weights models such as GLM-5.2 (max) and DeepSeek V4 Pro (max), with GLM-5.2 (max) scoring only ~90 Elo below Claude Opus 4.8 (max) for less than 25% of the cost
➤ Real-world complexity remains difficult for models: The top performer, Claude Fable 5, satisfies all rubric criteria on just 3% of AA-Briefcase tasks. On 31 of 91 tasks, no model scores above 50% on the rubric criteria
➤ Task difficulty scales with the number of required input files: For each rubric check, we identify the set of source files needed to pass. Across all models, pass rates fall as this file count increases, though top-tier models degrade less than weaker models
More details below in thread ⬇️
See 4 related tweets
- @teortaxesTex: GLM-5.1 came out 2 weeks before V4, and was already stronger. they don't have GLM-5 scores in any ca...
- @ClementDelangue: “Cost per task varies by ~800x across models tested: Claude Fable 5 leads the benchmark but costs mo...
- @TeksEdge: 🌱 Benchmark Evolved! 💼 Briefcase is a new benchmark from @ArtificialAnlys designed to test AI on rea...
- @Zai_org: RT @ArtificialAnlys: Announcing AA-Briefcase, the benchmark for the next era of agentic knowledge wo...
6. etnshow (Group Score: 144.3 | Individual: 35.2)
Cluster: 6 tweets | Engagement: 22 (Avg: 83) | Type: Tech
Co-Founder & CEO of @Doubleword_ Meryem (@MeryemArik9) says open-source AI is critical to building a stable long-term AI ecosystem and avoiding a future dominated by a handful of closed providers:
"We've believed from the very beginning that open-source models need to win if we're going to have a stable AI future."
"Otherwise, we risk descending into some kind of weird dystopia."
"We've spent the last three years building infrastructure for open-source model inference."
"What we care about is very high efficiency model inference: getting the most tokens and the most intelligence possible for a fixed dollar amount."\n\nQT @pmarca: Interesting. (Founder of https://t.co/PY9jRPdR6L, creator of GLM AI models.) https://t.co/cQFG1Q898Q
See 5 related tweets
- @etnshow: "Owning your AI and adopting open-source models is the single most important thing we can do."
@Mer...
- @zerohedge: "We created a monster" companies rein in AI usage as costs strain budgets
“Compute costs are now b...
- @ZeffMax: OpenAI President Greg Brockman: "the number of people using these agents is, like, the order of 10 m...
- @theinformation: Noam Shazeer, one of Google’s most important AI researchers, is joining OpenAI.
His move gives Ope...
- @zerohedge: Goldman predicted 120 quadrillion tokens monthly by 2030. It didn't say where. https://t.co/tI3Mwfl5...
7. BrianRoemmele (Group Score: 118.5 | Individual: 29.1)
Cluster: 5 tweets | Engagement: 54 (Avg: 309) | Type: Tech
“I can’t believe how this PsyOp Prompt works so well. I ran it on 53 news stories from the past and the output is crazy!”—X Subscriber
Thank you! It is a power tool. Be mindful of how you use it.\n\nQT @BrianRoemmele: The Psyops Identification Tool and a Universal Analysis AI Prompt is got folks really worried.
“Bigger and better data sets and powerful AI models are allowing us to spot previously undetectable patterns”.
Don’t do news without it. https://t.co/zkDOpUmdxD
See 4 related tweets
- @BrianRoemmele: The Psyops Identification Tool and a Universal Analysis AI Prompt is got folks really worried.
“Bi...
- @BrianRoemmele: BOOM!
JUST IN TIME FOR DISCLOSURE!
FOR MY SUBSCRIBERS!
A NEW PROMPT:
The Psyops Identification ...
- @BrianRoemmele: RT @BrianRoemmele: The Psyops Identification Tool and a Universal Analysis AI Prompt is got folks r...
- @BrianRoemmele: RT @BrianRoemmele: FOR MY SUBSCRIBERS!
A NEW PROMPT:
The Psyops Identification Tool and a Univers...
8. anshnanda (Group Score: 117.9 | Individual: 35.7)
Cluster: 4 tweets | Engagement: 187 (Avg: 32) | Type: Tech
In 12 months millions of people will be sharing, buying and selling AI agents. But building a platform that makes these agents customizable and sharable is not easy. But after 6 months, I think we figured it out. Announcement soon.
See 3 related tweets
- @sairahul1: Imagine you decided to build your first AI agent:
complete confusion, frustration no idea how t...
- @AnatoliKopadze: Sam Altman:
"We're going to see 10-person billion-dollar companies pretty soon"
"If I were 22 righ...
- @rileybrown: Someone should build Hermes that runs in the cloud, except more reliable, and you can set it up in 1...
9. VKazulkin (Group Score: 112.9 | Individual: 55.7)
Cluster: 3 tweets | Engagement: 1024 (Avg: 36) | Type: Tech
RT @eng_khairallah1: Google CEO, Sundar Pichai:
"If you don't learn to how to orchestrate agents now, you'll spend 2027 catching up to people who started today"
In 30 minutes he explains why the best engineers stopped writing code and started building agents.
Most people think building an agent requires an engineering degree.
It doesn't. It requires one guide and one afternoon.
Watch the interview, then save the exact setup below 👇
See 2 related tweets
- @mikenevermiss: Google CEO, Sundar Pichai:
"if you don't learn how to build and coordinate agents now, you'll spend...
- @Av1dlive: RT @Av1dlive: Google CEO, Sundar Pichai:
"If you don't learn how to orchestrate agents now, you'll ...
10. sydneyrunkle (Group Score: 109.3 | Individual: 27.9)
Cluster: 5 tweets | Engagement: 32 (Avg: 13) | Type: Tech
open models are having another moment! GLM-5.2 is remarkably good
i've been using it with deepagents code (dcode), our coding agent built on our open harness, which makes it super easy to take new models for a spin
open model 🤝 open harness https://t.co/z22bGZ73on
See 4 related tweets
- @LangChain_OSS: Getting started with deepagents code is super easy!
https://t.co/d9ElOOhMN2\n\nQT @sydneyrunkle: op...
- @aiedge_: GLM-5.2 is the most powerful open-sourced AI model ever.
The problem is, most people have no clue h...
- @AndrewCurran_: Rave reviews all over my timeline for GLM 5.2 the last 48 hours, open source is having an exceptiona...
- @rileybrown: Spent a lot of time using GLM 5.2.
I’ve always been skeptical of the Open Models, as they’ve never...
11. VaibhavSisinty (Group Score: 106.0 | Individual: 32.8)
Cluster: 4 tweets | Engagement: 98 (Avg: 133) | Type: Tech
This changes how you teach AI forever.
OpenAI just shipped a feature for Codex called Record and Replay. Instead of writing a prompt explaining every step of a task, you just do the task once while Codex watches.
It records your clicks, your steps, your preferences. Then it turns the whole thing into a reusable skill you can trigger anytime.
Here's what that looks like in practice.
A team publishes YouTube videos with a consistent process: pull metadata from a spreadsheet, match the right assets, add thumbnail and captions, upload as private, verify everything.
They showed Codex how to do it once. One recording.
Next video package comes in, they hand it to Codex, and it handles the entire upload on its own.
Right row on the spreadsheet, right metadata, right thumbnail, right captions, saved as private, verified.
And it works across everything: browser, computer use, connected plugins, or a combination.
Expense reports, pull requests, calendar setups, recurring downloads, whatever process you repeat.
The shift: we went from "tell the AI what to do" to "show the AI what to do."
That's a fundamentally different interface. And most people haven't realized it shipped yet.\n\nQT @OpenAIDevs: Show Codex a workflow once. Reuse it as a skill.
Record & Replay lets you show Codex a recurring task, like filing an expense report or submitting a time-off request.
Codex turns that demo into an inspectable, editable skill.
You control when recording starts and stops. https://t.co/UqSGaO7XUs
See 3 related tweets
- @WesRoth: OpenAI released Record & Replay for the Codex app, allowing users to turn demonstrated computer ...
- @cgtwts: OpenAI just wiped out countless startups with a random Thursday drop. https://t.co/0K1Jze6pBv\n\nQT ...
- @RoundtableSpace: OPENAI JUST GAVE CODEX A RECORD & REPLAY FEATURE THAT TURNS YOUR WORKFLOWS INTO REUSABLE AI SKIL...
12. chandrarsrikant (Group Score: 101.1 | Individual: 31.2)
Cluster: 6 tweets | Engagement: 973 (Avg: 573) | Type: Tech
🚨 EXCLUSIVE: Meta in talks to invest in Kunal Shah's Cred
Significantly, they've also discussed a complete acquisition too!!
Facebook parent Meta, over the past weeks, has held discussions to invest in Indian fintech Cred, at least six people aware of the discussions told Moneycontrol.
While the tech giant has explored investing tens of millions of dollars in Cred as primary capital, it has also evaluated other options, including a full acquisition of the company at a lower valuation and bringing founder Kunal Shah into the organisation in an operating role, sources told Moneycontrol.
“Which path Meta ultimately chooses remains unclear. What is evident, however, is that Cred is in need of fresh capital and Meta has expressed a willingness to back the company,” one of the people cited above said.
Break with @Goenka_Tushar1
See 5 related tweets
- @chandrarsrikant: Something big cooking in CRED..details ⏬⏬\n\nQT @Goenka_Tushar1: New: Meta has held talks to invest ...
- @chandrarsrikant: RT @moneycontrolcom: 🚨While the tech giant has explored investing tens of millions of dollars in Cre...
- @chandrarsrikant: RT @moneycontrolcom: #MCExclusive | 🚨Meta in talks to invest in Kunal Shah’s Cred
The proposed $4 b...
- @chandrarsrikant: RT @Goenka_Tushar1: New: Meta has held talks to invest in Cred at $4B, sources to @chandrarsrikant a...
- @chandrarsrikant: RT @CNBCTV18Live: #Meta in talks to invest in Kunal Shah’s #Cred
13. business (Group Score: 98.8 | Individual: 31.9)
Cluster: 4 tweets | Engagement: 243 (Avg: 70) | Type: Tech
Jio Platforms is gearing up for what could become India’s biggest IPO ever. The telecom giant serves nearly 525 million subscribers and counts Meta, Alphabet and KKR among its backers. Now, Ambani’s digital crown jewel is preparing for its next chapter.
@talktosanjai has more: https://t.co/asKEjUXDl7
See 3 related tweets
- @moneycontrolcom: #RILAGM2026 | 🚨Reliance AGM- Jio IPO, Retail IPO & AI Updates Expected 💸
Chairman Mukesh Ambani is ...
- @moneycontrolcom: #RILAGM | 🚨 Reliance Jio has filed its Draft Red Herring Prospectus (DRHP) for an IPO.
Announcing t...
- @moneycontrolcom: #RILAGM2026 | 🚨🚨 Reliance’s Jio Platforms to file DRHP today, Mukesh Ambani says IPO will lead 'next...
14. chandrarsrikant (Group Score: 97.9 | Individual: 30.1)
Cluster: 4 tweets | Engagement: 56 (Avg: 573) | Type: Tech
🚨 New details on Innovist payout for founders. It's a significantly higher number than what's been reported so far.
Innovist, L'Oreal deal to give founders Rs 1,800 crore payout; Falguni Nayar, Sauce VC, Amazon to get 10X returns
In all, the three founders stand to make Rs 1,700-1,800 crore in total by selling Innovist to L'Oreal. While they get immediate payouts of about Rs 700-800 crore, they stand to net another Rs 1,000-1,100 crore over a 3-4 year period once L'Oreal makes a full acquisition and buys 100% of the company, sources told Moneycontrol.
Early backers including consumer-focused venture capital fund Sauce VC, Nykaa founder Falguni Nayar's family office 72 Ventures LLP, Amazon, ICICI Venture and a slew of other investors are set to walk away with significant returns on their investment.
By @Goenka_Tushar1
See 3 related tweets
- @chandrarsrikant: For Innovist's founders and investors, the deal represents the culmination of years of conviction. h...
- @BhavyaDKumar: RT @Goenka_Tushar1: New: Innovist and L'Oreal Rs 4000 cr deal to put Rs 1800 cr in hands of three fo...
- @moneycontrolcom: #MCExclusive | 🚨Innovist, L’Oreal deal to give founders Rs 1,800 crore payout; Falguni Nayar, Sauce ...
15. 0x_kaize (Group Score: 95.1 | Individual: 36.8)
Cluster: 3 tweets | Engagement: 40 (Avg: 70) | Type: Tech
Cursor turned coding into vibes
Polsia just turned starting a company into vibes
watched Ben's new video:
he flew to San Diego to meet the actual humans building real businesses on top of an AI that does the coding, the support, the ads, the fundraising
ONE FOUNDER per company & zero employees
actually, you don't need a team to test a business idea anymore
you need an idea, a laptop, and the willingness to ship something imperfect at speed
every founder I know is going to watch this and quietly start a project tonight
the cost of starting a company just dropped to near zero\n\nQT @Bencera: I built Polsia into a $250M company in under 3 months.
Solo + AI. Zero employees.
Everyone asks me how I did it.
Introducing aisloP, a docu-series on how I build Polsia.
Episode 1: The Launch.
How I orchestrated the biggest Twitter launch of 2026. https://t.co/LqYvyZIRwq
See 2 related tweets
- @alex_prompter: $250M valuation. Headcount: one human.
The rest of the company is agents.
Code, ads, support, ship...
- @shiri_shh: "We're going to see ONE-person BILLION-dollar companies pretty soon.
If I were 22 right now, I'd fe...
16. gdgtify (Group Score: 92.5 | Individual: 32.0)
Cluster: 3 tweets | Engagement: 11 (Avg: 310) | Type: Tech
I love these hybrid image prompts. This one has a lot of potential.
A hybrid image that reads as [Subject A] up close and as [Subject B] from a distance or when squinting ::5 Frequency Logic: [Subject A] lives in the fine sharp detail; [Subject B] lives in the soft blurry low-frequency tones — infer two subjects whose shapes roughly align so both layers register ::5 Composition: Both subjects centered and same scale, edges aligned, neutral background ::3 Material Physics: Crisp high-detail layer overlaid on a soft tonal layer, balanced contrast ::2 Illumination: Flat even light so neither layer dominates ::1 Render Stack: Optical-illusion artwork, MIT hybrid-image study, fine-art photo composite ::2 Negative: [only one subject readable, misaligned subjects, too much detail drowning the far reading, harsh outlines, color clash, text] :: -1\n\nQT @Gdgtify: GPT can come up with some clever product designs. I'd definitely consider buying a couple of these if someone made them.
2x2 grid, do this for 4 dinosaurs, for a mug and 3 other home products: generate a clever [product_category] inspired by [subject]. rules: - do not merely decorate the product with subject - transform subject into functional industrial design that is feasible and world class designed - infer form_language from subject - infer affordance_map so subject features become usable product features - primary_function must remain clear and practical - product must feel manufacturable, ergonomic, stable, and desirable - the cleverness should come from semantic feature transfer: shape -> function, anatomy -> utility, symbolism -> affordance - preserve recognizability of subject while making it a believable real product - infer material_system and design_tone from product_category + subject - no hard-coded handles, spouts, lids, or controls unless logically inferred
See 2 related tweets
- @gdgtify: I asked GPT to design me 4 nature inspired home gadgets. A couple of them have potential, don't you ...
- @gdgtify: I know I do a lot of weird prompts but here is a readable one for custom characters.
Prompt: gene...
17. jerryjliu0 (Group Score: 91.5 | Individual: 35.6)
Cluster: 3 tweets | Engagement: 99 (Avg: 42) | Type: Tech
It's kind of crazy how well LiteParse does on markdown document parsing even compared against frontier VLMs - when it doesn't use VLMs or any AI/OCR models at all. It's pure code.
On ParseBench, it outperforms Qwen 3.5-9B / GLM-OCR.
There's still a gap vs. models like Gemma 4 and PaddleOCR-VL especially on dense visual outputs, but if your documents are text/table-heavy this gap closes rapidly.
Come check it out: it's the fastest document parser you can possibly use, and it's completely free/open-source.
Repo: https://t.co/JNER0mVcB8\n\nQT @jerryjliu0: We built the fastest PDF -> markdown parser in the world 🚀⚡️
AND it’s more accurate than any other open-source, model-free parser (pymupdf4llm, opendataloader, pdf-inspector, markitdown) on 3 standardized benchmarks: olmOCR0-bench, opendataloader-bench, ParseBench
Introducing LiteParse v2.1. The v2 base version was already the fastest document->text parser on the planet, and with this new release we’ve introduced markdown.
It is fully open-source (Apache 2.0) and free, is usable from CLI/Rust/Node/Python/WASM, and is also installable as a one-click agent skill.
Check it out: https://t.co/7oFImAZeb2
Come check out LiteParse: https://t.co/JNER0mVcB8
See 2 related tweets
- @llama_index: RT @jerryjliu0: We built the fastest PDF -> markdown parser in the world 🚀⚡️
AND it’s more accurate...
- @itsclelia: RT @jerryjliu0: It's kind of crazy how well LiteParse does on markdown document parsing even compare...
18. StockSavvyShay (Group Score: 90.2 | Individual: 32.0)
Cluster: 3 tweets | Engagement: 389 (Avg: 593) | Type: Tech
AI memory is turning into one of the biggest profit pools in the entire semiconductor stack.
HBM has genuinely converted memory into qualification-gated AI infrastructure with 1T of operating income by 2029. https://t.co/MoU5xIUC3h
See 2 related tweets
- @StockMKTNewz: The memory companies could be about to make $3 Trillion of profits this decade https://t.co/4m0VLzka...
- @wallstengine: Samsung, SK Hynix and Micron could generate nearly $3T of cumulative operating income from 2026-2029...
19. BrianRoemmele (Group Score: 88.4 | Individual: 29.6)
Cluster: 4 tweets | Engagement: 79 (Avg: 309) | Type: Tech
My video insights from the Midjourney Medical Full Body ultrasound transducer system.
Read the member only deep dive and investment guide, some folks are up 57% in a few days.
Join us: https://t.co/Q7MGTNckYY https://t.co/wp1KwaA0Xv\n\nQT @BrianRoemmele: Midjourney launched a revolutionary full-body AI medical scanner that aims to make preventive healthcare accessible transforming health monitoring to routine home health improving early disease detection, enhancing long-term health outcomes.
See 3 related tweets
- @BrianRoemmele: RT @BrianRoemmele: My video insights from the Midjourney Medical Full Body ultrasound transducer sys...
- @BrianRoemmele: RT @BrianRoemmele: Midjourney launched a revolutionary full-body AI medical scanner that aims to mak...
- @WesRoth: Midjourney published a technical breakdown of its experimental full-body ultrasound system, called t...
20. noahzweben (Group Score: 84.0 | Individual: 31.3)
Cluster: 3 tweets | Engagement: 106 (Avg: 106) | Type: Tech
I use these almost every day internally to share information with my team\n\nQT @claudeai: New in Claude Code: Artifacts.
Interactive pages built from your session, like a PR walkthrough or a living project dashboard, shared with your team at a private link.
Available in beta on Team and Enterprise plans. https://t.co/0NX9gNCaAs
See 2 related tweets
- @owenbjennings: been great to have a model-agnostic version of this @blocks since august last year [ blockcell ]
u...
- @realsanketp: RT @claudeai: New in Claude Code: Artifacts.
Interactive pages built from your session, like a PR w...