@mweinbach Max Weinbach

Max Weinbach posts on X about apple, this is, ai, macbook the most. They currently have [-------] followers and [---] posts still getting attention that total [-------] engagements in the last [--] hours.

Engagements: [-------] #

Mentions: [--] #

Followers: [-------] #

CreatorRank: [-------] #

Social Influence

Social category influence technology brands 23.04% stocks 12.57% products #2022 automotive brands 4.71% gaming 3.66% finance 2.09% social networks 1.57% events #513 financial services 0.52% luxury brands 0.52%

Social topic influence apple #2479, this is 7.85%, ai 7.33%, macbook #91, open ai #786, iphone #1550, overwatch #609, tesla 3.66%, if you 3.66%, $googl #1448

Top accounts mentioned or mentioned by @itsnoahd @sigkitten @snazzylabs @sawyermerritt @luftkopptim @elaina43114880 @aurexav @tesla @_apbright @llmjunky @nicdunz @vadimyuryev @apbright @davidpiccinini2 @exoticspice101 @tonywhitehawk @benbajarin @providerproto @nerdymomocat @lukewat48544961

Top assets mentioned Tesla, Inc. (TSLA) Alphabet Inc Class A (GOOGL) Dell Technologies, Inc. (DELL) Microsoft Corp. (MSFT)

Top Social Posts

Top posts by engagements in the last [--] hours

"I was looking at why battery life wasn't amazing on a laptop I'm testing so I was trying to figure out why Turns out in 120hz dynamic refresh rate the OLED panel was using as much or more power than the SoC itself"
X Link 2026-02-17T19:10Z 284.3K followers, 13.2K engagements

"Currently tracking a system to nearly 2x battery life by setting the display to 60hz from 120hz dynamic Insane"
X Link 2026-02-18T00:29Z 284.3K followers, 53.1K engagements

"Got my first panther lake laptop Galaxy Book [--] Pro"
X Link 2026-02-05T21:59Z 284.3K followers, 108.6K engagements

"@tonywhitehawk Samsung"
X Link 2026-02-18T00:47Z 284.3K followers, [----] engagements

"Xperia [--] II specs and images"
X Link 2020-02-23T14:22Z 284.3K followers, [----] engagements

"This is from Apple's State of the Union The local model is a 3B parameter SLM that uses adapters trained for each specific feature. Diffusion model does the same thing adapter for each style. Anything running locally or Apple's Secure Cloud is an Apple model not OpenAI"
X Link 2024-06-10T21:21Z 284.2K followers, 2.7M engagements

"Apple just made the MacBook Air even better: M4 chip Gorgeous new Sky Blue color 18-hour battery life Better camera (finally) And $100 cheaper ($999 now) for 16GB/256GB Its available March 12"
X Link 2025-03-05T14:02Z 284.3K followers, 1.4M engagements

"The M2 MacBook Air is going for $699 right now absurd value for it. 256GB storage 16GB RAM. https://howl.link/1c36063tdrcv6 https://howl.link/1c36063tdrcv6"
X Link 2025-03-09T14:29Z 284.2K followers, 452K engagements

"768GB RAM [---] GPU cores [--] CPU cores 12TB storage Welcome to my desk 512GB RAM M3 Ultra Mac Studio (this is why I got a thunderbolt [--] cable)"
X Link 2025-03-14T15:02Z 284.2K followers, 777.6K engagements

"lol looks like Mark Rober edited in a Pixel image over the iPhone he was recording with. You can see the iPhone in the reflection"
X Link 2025-03-16T17:46Z 284.3K followers, 1.7M engagements

"Just updated my Pixel [--] Pro XL got the new Material [--] Expressive UI"
X Link 2025-05-20T18:26Z 284.2K followers, 328.1K engagements

"The M2 MacBook Air is $699 at Best Buy This is cheaper than some Chromebooks and is going to be the best laptop period at its price"
X Link 2025-08-09T00:38Z 284.2K followers, 283K engagements

"Half of Mac buyers were new to Mac. Grew 13% year over year due to new MacBook Air and lower starting price"
X Link 2025-10-30T21:20Z 284.2K followers, 2.9M engagements

"Reading a JP Morgan AI CAPEX report and this is such a great way to put AI ROI into perspective "to drive a 10% return on our modeled AI investments through [----] would require $650 billion of annual revenue into perpetuity. which equates to $34.72/month from every current iPhone user""
X Link 2025-11-10T15:59Z 284.2K followers, 755.8K engagements

"So the $40k Mac studio cluster I have with the 1.5TB of memory. It's to run the new @exolabs [---] which launches today It lets you connect these devices with Thunderbolt [--] RDMA and use tensor parallelism for running models like Kimi K2 K2 Thinking runs at [--] tok/s @ 280W"
X Link 2025-12-18T14:46Z 284.2K followers, 249.9K engagements

"Intel Wildcat Lake in the flesh"
X Link 2026-01-06T01:08Z 284.3K followers, 51.9K engagements

"Now that ChatGPT has synced the past decade of my apple health data heres Claude with Health vs ChatGPT Health"
X Link 2026-01-18T15:57Z 284.3K followers, 53.7K engagements

"@pvncher cerebras systems also support up to 120T parameter models with MemoryX modules https://www.cerebras.ai/blog/announcing-the-cerebras-architecture-for-extreme-scale-ai https://www.cerebras.ai/blog/announcing-the-cerebras-architecture-for-extreme-scale-ai"
X Link 2026-01-19T03:40Z 284.2K followers, [---] engagements

"Anyone paying attention to Apple should not be surprised by their earnings And any call for them missing out on AI or being late is just stupid. It'll come it obviously is this year and being "late" didn't hit them at all and likely helped with the earnings we see"
X Link 2026-01-29T21:55Z 284K followers, 10.8K engagements

"M5 Apple Silicon is seeing some pickup for on-device AI: "AstraZeneca is rolling out over [----] M5 powered iPad pros to its pharmaceutical sales team to take advantage of AI capabilities including Apple intelligence while meeting with clinicians daily." - Kevan Parekh"
X Link 2026-01-29T22:22Z 284.2K followers, 27.5K engagements

"Local Kimi K2.5 the model FINALLY downloaded"
X Link 2026-02-01T02:23Z 284.2K followers, 109K engagements

"@teslascope just signed up gonna use this for sure"
X Link 2026-02-02T00:25Z 284K followers, [---] engagements

"Maaaan I just want to take a screenshot whats the difference why are there two"
X Link 2026-02-02T04:17Z 284.2K followers, 457.7K engagements

"And back to the OnePlus 15"
X Link 2026-02-02T14:35Z 284.2K followers, 27.7K engagements

"I feel like the Codex app or possible GPT-5.3 arent the most exciting things coming from OpenAI in the next [--] days"
X Link 2026-02-02T19:46Z 284.1K followers, 29.3K engagements

"Would HIGHLY recommend connecting the Malwarebytes app to ChatGPT"
X Link 2026-02-03T01:48Z 284K followers, 57K engagements

"setting xcode codex reasoning to xhigh in the config.toml helps a lot it defaults to medium"
X Link 2026-02-03T21:47Z 284.1K followers, 12.5K engagements

"oh sweet you can replace the xcode Codex binary with the latest version and make it use subagents i wouldn't recommend it since the xcode ui doesn't support it and it's a bit funky but you can do it"
X Link 2026-02-03T22:10Z 284.2K followers, [----] engagements

"With iOS [----] RC finally coming out that means were close to [----] and the new Siri"
X Link 2026-02-04T19:16Z 284.2K followers, 27.7K engagements

"Just to help set expectations new Siri adds [--] new features/capabilities [--]. On screen awareness [--]. person context [--]. taking action in/across apps Maybe they surprise with more but this is all theyve said they are definitely working on Chatbot version is apparently iOS [--] @mweinbach Its not going to be great. @mweinbach Its not going to be great"
X Link 2026-02-04T19:32Z 284.1K followers, 109.5K engagements

"750M monthly active users for Gemini Theyve almost caught up to ChatGPT thats insanely impressive Our Q4/FY25 results are in. Thanks to our partners & employees it was a tremendous quarter exceeding $400B in annual revenue for the first time. Our full AI stack is fueling our progress and Gemini [--] adoption has been faster than any other model in our history. Were really https://t.co/UbcQPFRGkr Our Q4/FY25 results are in. Thanks to our partners & employees it was a tremendous quarter exceeding $400B in annual revenue for the first time. Our full AI stack is fueling our progress and Gemini 3"
X Link 2026-02-04T21:32Z 284.3K followers, 121.2K engagements

"@gosu984 Where is that number from because I dont think they have half the worlds population as monthly active users"
X Link 2026-02-04T21:37Z 284.2K followers, [----] engagements

"@Marks_Tech I dont think its that especially since this includes web and iOS Just because its installed it doesnt mean its used"
X Link 2026-02-04T22:27Z 284.2K followers, [----] engagements

"Throwing Gemini [--] Flash in Claude Code in xcode for doing UI design with the preview tool is pretty great"
X Link 2026-02-04T23:06Z 284.2K followers, [----] engagements

"I really can't get over how good the Claude ads are It's not even the advertising part but the ghoulish awful way the AI models talk and communicate. It is so awful. It's like nails on a chalkboard"
X Link 2026-02-05T02:06Z 284.2K followers, 34.8K engagements

"GPT [---] Codex is now in the CLI and app"
X Link 2026-02-05T18:05Z 284.1K followers, 18.6K engagements

"I can't believe we got GPT [---] Codex and Claude Opus [---] in a [--] minute span"
X Link 2026-02-05T18:07Z 284.2K followers, 26.8K engagements

"OH MY GOD ANTHROPIC DID IT IT DOESN'T FORCE BULLET POINTS"
X Link 2026-02-05T19:42Z 284.2K followers, 44.1K engagements

"Ok Im finding GPT [---] Codex is the best coding model for anything technical but Opus [---] can do some amazing UI work (a given) and is better with subagents But Opus [---] in Claude Cowork is night and day better. Best version of a generalized agent hands down. Game changer imo"
X Link 2026-02-05T19:54Z 284.2K followers, 29K engagements

"OH I must say this Panther Lake laptop is the first Windows laptop I've used in a long time that just feels fast The Snapdragon X Elite laptops all feel like this but this is the first x86 laptop that does as well I think it's part scheduler and part sheer power"
X Link 2026-02-06T00:26Z 284.2K followers, 16.3K engagements

"ASUS ROG Zephyrus G16 (the [----] one) is such a good gaming laptop wow"
X Link 2026-02-06T17:09Z 284.2K followers, [----] engagements

"Being able to control CPU core configuration and enable/disable P and E cores on Arrow Lake is cool This is a laptop too Asus Zephyrus G16 with Intel Core Ultra [--] 285H"
X Link 2026-02-06T22:59Z 284.3K followers, [----] engagements

"Its 2.5x faster at 6x the price I have little doubt in my mind almost all enterprise customers of Opus pay for this And overnight Anthropic has likely increased ARR 25%+ Our teams have been building with a 2.5x-faster version of Claude Opus [---]. Were now making it available as an early experiment via Claude Code and our API. Our teams have been building with a 2.5x-faster version of Claude Opus [---]. Were now making it available as an early experiment via Claude Code and our API"
X Link 2026-02-07T19:20Z 284.1K followers, 70K engagements

"On a laptop I think when looking at displays my ranking is Tandem OLED MiniLED LCD OLED Standard OLED uses way too much power. Id rather get 25% more battery life and take a good LCD or Mini LED panel But ofc would take tandem OLED above all else"
X Link 2026-02-07T22:26Z 284.3K followers, 99.4K engagements

"Let's see how claude agent code teams are with gpt [---] codex as the model"
X Link 2026-02-08T00:24Z 284.3K followers, 38.6K engagements

"If it uses A18 Pro those chips are packaged with TSMC InfoPoP which means the memory is physically stacked on top of the SoC Likely means 8GB will be the only option New midrange MacBook will apparently come with 8GB ram https://t.co/cQlq6xyBQ5 New midrange MacBook will apparently come with 8GB ram https://t.co/cQlq6xyBQ5"
X Link 2026-02-08T14:02Z 284.2K followers, 188.6K engagements

"@_philschmid Yea I just want it to work better first so I dont have people complaining"
X Link 2026-02-08T14:30Z 284.1K followers, [----] engagements

"aaand an ios version so you can access your Android phone's texts on an iPhone Had GPT [---] Codex make a native app for Google Messages on macOS Few things to fix up but can send/receive messages (including RCS) on my Pixel from my Mac https://t.co/BIKiKqG4pC Had GPT [---] Codex make a native app for Google Messages on macOS Few things to fix up but can send/receive messages (including RCS) on my Pixel from my Mac https://t.co/BIKiKqG4pC"
X Link 2026-02-08T17:23Z 284K followers, 21.5K engagements

"anyone else get the merch drop"
X Link 2026-02-09T00:17Z 284.1K followers, 19.4K engagements

"as have i but there's a lot more going on here. okay i have been alerted by multiple openai people that this entire video and concept is fake including the reddit post that it originated from. please ignore. okay i have been alerted by multiple openai people that this entire video and concept is fake including the reddit post that it originated from. please ignore"
X Link 2026-02-09T03:20Z 284K followers, 16K engagements

"@Kaligula_Khan @signulll here's the problem got these a week ago"
X Link 2026-02-09T03:27Z 284.1K followers, 14.8K engagements

"This is not OpenAI so the question is what is this Alexander Skarsgrd stars in a mysterious #SuperBowl commercial for a product called 'dime' https://t.co/xlgtsjPqIB Alexander Skarsgrd stars in a mysterious #SuperBowl commercial for a product called 'dime' https://t.co/xlgtsjPqIB"
X Link 2026-02-09T04:55Z 284.2K followers, 327.7K engagements

"That Jony Ive Ferrari console looks so good"
X Link 2026-02-09T15:44Z 284.1K followers, 11.2K engagements

"@ajrgd https://consumer.huawei.com/bh/headphones/freeclip2/ https://consumer.huawei.com/bh/headphones/freeclip2/"
X Link 2026-02-09T15:46Z 284.3K followers, [----] engagements

"Snapdragon X2 Elite seems like we're getting absurd performance per watt in Windows machines I can not wait to try it Snapdragon X2 Elite benches the Apple M5 and asks You good Thanks to Hardware Canucks we're finally getting to look at real Snapdragon X2 Elite benchmarks (early drivers) versus Apple's M5 (and Intel and AMD) and this is the first time were seeing what Qualcomms nextgen https://t.co/d8C1qYDwY0 Snapdragon X2 Elite benches the Apple M5 and asks You good Thanks to Hardware Canucks we're finally getting to look at real Snapdragon X2 Elite benchmarks (early drivers) versus Apple's"
X Link 2026-02-09T15:48Z 284.3K followers, 10.1K engagements

"Because its relevant the gaming performance on Panther Lake is good enough to play Overwatch at a stable 120fps on a thin and light laptop This is so nice"
X Link 2026-02-10T00:14Z 284.2K followers, 13.3K engagements

"My cowork agent is coming along well just testing it out I have an interesting idea for making sure it doesn't delete your entire computer too"
X Link 2026-02-10T01:24Z 284.1K followers, [----] engagements

"and because it's that time Panther Lake geekbench score was in Windows more performance mode but extremely performant"
X Link 2026-02-10T01:57Z 284.2K followers, 10.4K engagements

"Woah ChatGPT will add support for importing skills to your library I just had it create a skill for me that I could use in Codex and got this popup in the chat"
X Link 2026-02-10T02:52Z 284.3K followers, 83.8K engagements

"Rivian is using 5G radios too Likely Qualcomm Tesla has this 5G modems in their cars as well but they are not enabled. Rivian enabling it is very nice QoL R2 has a 5G radio https://t.co/A9nk1e91hK R2 has a 5G radio https://t.co/A9nk1e91hK"
X Link 2026-02-10T15:25Z 284.3K followers, 24.8K engagements

"@thsottiaux Right now Im noticing sometimes it stops updating the UI even though codex is still working and the only way Im able to see it completed the task is quitting and opening the app again That would be nice"
X Link 2026-02-10T15:28Z 284.1K followers, [---] engagements

"@BenBajarin The entire market is at production capacity and theres still a surplus in demand AI revenue is entirely limited by TSMC/memory manufacturers supply capacity"
X Link 2026-02-10T15:30Z 284.1K followers, [----] engagements

"Ive somehow had this for like a week and its very good and a massive improvement Deep research in ChatGPT is now powered by GPT-5.2. Rolling out starting today with more improvements. https://t.co/LdgoWlucuE Deep research in ChatGPT is now powered by GPT-5.2. Rolling out starting today with more improvements. https://t.co/LdgoWlucuE"
X Link 2026-02-10T19:18Z 284.1K followers, 11.1K engagements

"@nicdunz It was based on o3 originally and they never said they updated it so Id assume its still o3"
X Link 2026-02-10T19:30Z 284.1K followers, [----] engagements

"@zeugma95031605 Its definitely better"
X Link 2026-02-10T19:43Z 284.1K followers, [---] engagements

"@rllytryingg @nicdunz Continuously being invested in was that agent/deep researcher merger I think the only update to the model was for Atlas"
X Link 2026-02-10T20:08Z 284.1K followers, [----] engagements

"This thing is sick Haptic trackpads and haptic mice who would have thought. Genuinely cool technology. I did purchase one instantly This is absolute Peak Gaming Mouse Here's our Logitech PRO X2 SUPERSTRIKE Review https://t.co/fz7eFmDhfH https://t.co/rNL3ka3NE1 This is absolute Peak Gaming Mouse Here's our Logitech PRO X2 SUPERSTRIKE Review https://t.co/fz7eFmDhfH https://t.co/rNL3ka3NE1"
X Link 2026-02-10T20:45Z 284.1K followers, [----] engagements

"I think Domina is my favorite of the new Overwatch heroes to play"
X Link 2026-02-10T22:41Z 284.1K followers, [----] engagements

"Samsung finally announced unpacked Feb 25th in San Francisco https://youtu.be/l8iOk6kkuuQsi=a6cTm6GiizQMAf3h https://youtu.be/l8iOk6kkuuQsi=a6cTm6GiizQMAf3h"
X Link 2026-02-10T23:09Z 284.1K followers, [----] engagements

"Maybe an interesting benchmark for silicon in Overwatch is how long it takes to render a highlight"
X Link 2026-02-11T02:54Z 284.1K followers, [----] engagements

"@AIadventure3 @DicksonPau @apples_jimmy GPT [---] Codex was designed for training and inference on GB200s Codex models are an iteration of GPT 5._ base models That means GPT [---] is also inference on GB200s and GB200s are very limited still"
X Link 2026-02-11T03:20Z 284.1K followers, [--] engagements

"Can't wait for GLM-5 to hit the API looking. better than Opus [---] Didn't expect this to happen. Gotta see if it's slop or has taste tho"
X Link 2026-02-11T13:42Z 284.2K followers, 14.7K engagements

"MiniMax M2.5 GLM-5. big day for open models What's next"
X Link 2026-02-11T13:49Z 284.2K followers, [----] engagements

"@rudrank oh no i didn't gave up on trying the stealth models lol did you any good"
X Link 2026-02-11T14:05Z 284.1K followers, [---] engagements

"@Elaina43114880 DeepSeek has been behind for a while now compared to the other chinese models DeepSeek has great research and model architecture I mean everyone uses them but their final product is worse. The others aren't possible without DeepSeek but DeepSeek itself isn't as good"
X Link 2026-02-11T14:11Z 284.1K followers, [---] engagements

"Technical report for GLM [--] is out it looks really good Looks nearly as good as Opus [---] actually. It's much larger than GLM [---] though at 774B total 40B active https://z.ai/blog/glm-5 https://z.ai/blog/glm-5"
X Link 2026-02-11T17:10Z 284.3K followers, 13.7K engagements

"@zephyr_z9 GLM [---] was what 22T DSA is nice tho should absolutely nuke available memory"
X Link 2026-02-11T17:29Z 284.3K followers, [---] engagements

"@zephyr_z9 200k output tokens is 128k because it can reason more need to see more 400k like gpt models"
X Link 2026-02-11T17:32Z 284.1K followers, [---] engagements

"@itsnoahd Same sorta I love Claude models but refuse to pay API pricing so these models are about as close as I can get without paying $25/1M tokens lol"
X Link 2026-02-11T17:40Z 284.1K followers, [---] engagements

"WEIGHTS ARE LIVE https://huggingface.co/zai-org/GLM-5 https://huggingface.co/zai-org/GLM-5"
X Link 2026-02-11T17:48Z 284.2K followers, [----] engagements

"Let's play a game of guess the model"
X Link 2026-02-11T18:16Z 284.2K followers, [----] engagements

"And now we begin testing of panther lake on our freshly updated battery benchmark"
X Link 2026-02-11T19:56Z 284.3K followers, 16.5K engagements

"@MichaelDell Cant wait to get mine"
X Link 2026-02-11T20:00Z 284.2K followers, [----] engagements

"Gurman has been reporting 26.4/26.5 for a while and that was likely the plan/goal from Apple But its worth noting Apple just said itll come later this year. So while it may be delayed internally it hasnt been delayed any further from Apples public timelines. New Siri Runs Into Problems Features Could Be Pushed to iOS [----] and iOS [--] https://t.co/XkoSb7FCer https://t.co/c1YdealYHO New Siri Runs Into Problems Features Could Be Pushed to iOS [----] and iOS [--] https://t.co/XkoSb7FCer https://t.co/c1YdealYHO"
X Link 2026-02-11T20:30Z 284.2K followers, 35.6K engagements

"@SnazzyLabs This doesnt surprise me basically it was too slow. Private cloud compute on older Apple silicon could be rough until their own accelerators and TPUs later on this year. Curious what theyll do. I hope its connect Apple silicon to the TPUs as the host computer"
X Link 2026-02-11T20:31Z 284.2K followers, [----] engagements

"Apple should just use GLM [--] to run Siri"
X Link 2026-02-11T20:32Z 284.2K followers, 11.3K engagements

"@scaling01 That makes sense. [--] Flash is apparently 1.2T according to Bloomberg (Apple Intelligence model)"
X Link 2026-02-11T20:33Z 284.2K followers, [---] engagements

"This is great because the show is so good and world is so interesting I want as much as possible of Severance Apple has acquired Severance (including the IP and all the rights) from production company Fifth Studio. The series is set to run for at least [--] seasons and could expand with prequels spinoffs and foreign versions. (https://t.co/6tRLvkEKLp) https://t.co/jBZ4XPIcIY Apple has acquired Severance (including the IP and all the rights) from production company Fifth Studio. The series is set to run for at least [--] seasons and could expand with prequels spinoffs and foreign versions."
X Link 2026-02-11T23:04Z 284.3K followers, 16.9K engagements

"Got a second Panther Lake system in. Asus ZenBook Duo This time Intel Core Ultra X9 the highest end SKU"
X Link 2026-02-11T23:43Z 284.2K followers, [----] engagements

"oh my god it literally burned through $40 worth of credits in [--] minutes and I didn't even notice holy shit"
X Link 2026-02-12T00:45Z 284.2K followers, [----] engagements

"@SIGKITTEN memory memory codex explain it it summarizes every chat and throws it into a memory file. but if you have a lot of previous chats. https://github.com/openai/codex/commit/53c1818d2982f27ee590bd6aadf6cf804d1491f0 https://github.com/openai/codex/commit/53c1818d2982f27ee590bd6aadf6cf804d1491f0"
X Link 2026-02-12T00:55Z 284.2K followers, [----] engagements

"@itsnoahd make sure to set your model to gpt-5.1-codex-mini before hand"
X Link 2026-02-12T00:56Z 284.2K followers, [----] engagements

"@itsnoahd it will read through every conversation you've ever had with codex and just burn through all your tokens [---] codex mini will 4x the available tokens for this"
X Link 2026-02-12T00:58Z 284.2K followers, [---] engagements

"@itsnoahd it's in development so i assume they have plans to not do this when it's actually released but yea i didn't realize lol it used a bunch of the credits i bought before i even realized"
X Link 2026-02-12T01:00Z 284.1K followers, [---] engagements

"@providerproto @SIGKITTEN well the app uses the cli so yea it's just codex looking at it's own open source repo"
X Link 2026-02-12T04:26Z 284.2K followers, [---] engagements

"@aurexav let's hope"
X Link 2026-02-12T05:22Z 284.2K followers, [----] engagements

"I'm trying to grab some fps and frame numbers from Overwatch on a couple systems and it's so funny that with DX12 fps goes down. Even enabling DLSS/FSR/XeSS with DX12 lowers fps below native rendering on DX11"
X Link 2026-02-12T15:07Z 284.2K followers, [----] engagements

"GLM [--] is available on Apple Silicon with MLX Gonna download and get it running on my cluster https://huggingface.co/mlx-community/GLM-5-4bit https://huggingface.co/mlx-community/GLM-5-4bit"
X Link 2026-02-12T16:37Z 284.3K followers, 22.6K engagements

"GPT [---] Codex Spark Its a smaller version of GPT [---] Codex running at over [----] tokens per second on Cerebras hardware GPT [---] Codex was trained for GB200 on GB200. I wonder what this is Maybe GPT [---] Codex that used Cerebras Reap GPT-5.3-Codex-Spark is now in research preview. You can just build thingsfaster. https://t.co/85LzDOgcQj GPT-5.3-Codex-Spark is now in research preview. You can just build thingsfaster. https://t.co/85LzDOgcQj"
X Link 2026-02-12T18:10Z 284.2K followers, 17.8K engagements

"Codex Spark was trained on GPUs for Cerebras hardware but OpenAI added support to their inference framework for Cerebras meaning theyre reading to load future models onto it too GPUs are still foundational for inference and training though"
X Link 2026-02-12T18:13Z 284.3K followers, 21.1K engagements

"@SteveMoser Yup Better than [---] codex mini"
X Link 2026-02-12T18:54Z 284.2K followers, [---] engagements

"@Alex_Intel_ Yup I wonder what happens to cost as more CPX style prefill machines hit the market with LPDDR memory. Should drive cost per token down"
X Link 2026-02-12T19:23Z 284.2K followers, [--] engagements

"@User__Ams007 ASUS sent it to me to review"
X Link 2026-02-12T22:15Z 284.2K followers, [--] engagements

"@nerdymomocat Mac version was denied but iOS version is up Ill try to fix the Mac version soon https://testflight.apple.com/join/hngdk6tR https://testflight.apple.com/join/hngdk6tR"
X Link 2026-02-12T22:32Z 284.2K followers, [----] engagements

"@alexgeorge23 Just because of the wafer scale I bet theyre running the full model on cereus with MTP and doing all the pre-fill on GPUs Idk if you could hit [----] even in one rack with both"
X Link 2026-02-13T00:57Z 284.2K followers, [--] engagements

"@nerdymomocat I def could Let make just make codex package and throw it into the repo"
X Link 2026-02-13T01:01Z 284.2K followers, [---] engagements

"How my Cowork agent built by GPT-5.3-Codex is going so far UI and how the agent works/responds needs work B U T it's going well Gemini [--] Flash also needs DCF help"
X Link 2026-02-13T03:20Z 284.3K followers, [----] engagements

"If you want access to the Google Messages mac sync here ya go https://github.com/mweinbach/SwiftGMessages/releases/tag/v1.0.0 https://github.com/mweinbach/SwiftGMessages/releases/tag/v1.0.0"
X Link 2026-02-13T03:43Z 284.2K followers, [----] engagements

"GPT-5.3-Codex-Spark is such a good model to play with You really can't tell it's smaller than GPT-5.3-Codex"
X Link 2026-02-13T04:33Z 284.3K followers, 14.3K engagements

"They must be freeing up compute for GPT-5.3 next week or something Tomorrow at 10am PT legacy models (GPT-5 GPT-4o GPT-4.1 GPT-4.1 mini and OpenAI o4-mini) will be deprecated in ChatGPT. https://t.co/RJioBsLY6D Tomorrow at 10am PT legacy models (GPT-5 GPT-4o GPT-4.1 GPT-4.1 mini and OpenAI o4-mini) will be deprecated in ChatGPT. https://t.co/RJioBsLY6D"
X Link 2026-02-13T04:53Z 284.2K followers, 40.8K engagements

"@M0_R4_H4 openai said they'll use it for larger models later"
X Link 2026-02-13T12:24Z 284.2K followers, [--] engagements

"Ok yea anything that is more than a basic task is worse on spark even if its faster. It ends up taking like 5x longer with worse work because of the number of compactions. Its good for just basic stuff. GPT-5.3-Codex-Spark is such a good model to play with You really can't tell it's smaller than GPT-5.3-Codex GPT-5.3-Codex-Spark is such a good model to play with You really can't tell it's smaller than GPT-5.3-Codex"
X Link 2026-02-13T15:12Z 284.2K followers, [----] engagements

"@awnihannun Have you seen transformer labs Similar idea and super built out https://transformerlab.ai/ https://transformerlab.ai/"
X Link 2026-02-13T15:29Z 284.2K followers, [---] engagements

"Jeez M3 Ultra Mac Studio estimates dates are now end of March early April I wonder if this is a Mac factory thing TSMC supply chain or memory supply chain thing Or if M5 Ultra is coming soon-ish People are waking up. Mac Studio hardware 54+ days out.FML https://t.co/KOdl0yqvfs People are waking up. Mac Studio hardware 54+ days out.FML https://t.co/KOdl0yqvfs"
X Link 2026-02-13T15:56Z 284.3K followers, 22.8K engagements

"This community note is hilarious"
X Link 2026-02-13T15:58Z 284.3K followers, [----] engagements

"Who would be interested in seeing a battery test of a Dell XPS with a tandem oled vs Dell XPS with LCD display"
X Link 2026-02-13T16:32Z 284.3K followers, 12.7K engagements

"The overarching test/thing I want to see is around the PC OEM side as margin compresses and BOM goes up where can you offer price reductions with a net equal or positive end user value Who would be interested in seeing a battery test of a Dell XPS with a tandem oled vs Dell XPS with LCD display Who would be interested in seeing a battery test of a Dell XPS with a tandem oled vs Dell XPS with LCD display"
X Link 2026-02-13T17:22Z 284.3K followers, [----] engagements

"@VadimYuryev @Hovscorpion12 Yea thats what I meant for products with it but idk I feel maybe May to get ahead of WWDC But also depends on supply chain Id imagine"
X Link 2026-02-13T17:27Z 284.3K followers, [---] engagements

"One of my worlds made it here :) You dreamed it. Project Genie built it. Explore some of our favorite worlds. 🌍 https://t.co/jifC8gw5nJ You dreamed it. Project Genie built it. Explore some of our favorite worlds. 🌍 https://t.co/jifC8gw5nJ"
X Link 2026-02-13T18:24Z 284.3K followers, 10.8K engagements

"Goodbye bad models Only the best shall remain"
X Link 2026-02-13T18:24Z 284.3K followers, 18.9K engagements

"@VictorTaelin Opus plan mode copy/paste prompt to codex. Honestly could probably automate this with an MCP or hook"
X Link 2026-02-13T19:07Z 284.2K followers, [---] engagements

"@nicdunz Technically its gpt-realtime and gpt-realtime-mini"
X Link 2026-02-13T22:16Z 284.2K followers, [----] engagements

"First time being paid by a defense contractor. Feels good"
X Link 2026-02-14T00:21Z 284.3K followers, 12K engagements

"@ivanfioravanti I didn't look too deeply into the new arch is it more attention/compute heavy"
X Link 2026-02-14T17:09Z 284.3K followers, [----] engagements

"What's funny about gpt [---] codex spark is the slowest part of the agent loop is my computer processing the tool calls"
X Link 2026-02-14T17:15Z 284.3K followers, [----] engagements

"this is llm batching they do say tokens are faster on Pro (and you def notice in practice) and fwiw Pro and Plus are slower than API so if you were to scale that the tokens from those two would be cheaper then an API token. Plus would probably be 50% cheaper and Pro probably 25% cheaper. This is why these services are profitable to run imo And i'd think about it like this if you have $32 in GPU hours (16 GPUs at $2 per hour each) in those hours we can generate at a batch size of [---] or [--]. With [--] each user will get something like [---] token/s. That's 3.2k tokens per second. At [---] it may be"
X Link 2026-02-15T00:21Z 284.3K followers, [---] engagements

"I'm enjoying GLM-5 a lot for quick UI/UX stuff in OpenCode Also the OpenCode desktop app looks WAY better than the last time I used it it's really nice"
X Link 2026-02-15T18:29Z 284.3K followers, [----] engagements

"@michael_k_woods @SnazzyLabs @SawyerMerritt @Tesla this has to be an ai reply"
X Link 2026-02-15T18:33Z 284.3K followers, [---] engagements

"if it's not. tesla includes all ADAS hardware in the cars they want to sell the subscription to you Their options to hit a higher ARR and sub uptake rate is lower prices but they also need to use subscriptions to offset lost rev other places which is higher prices the option they'll likely do is $250-300+ for unsupervised and $50-100 for supervised but still include the hardware in every car https://twitter.com/i/web/status/2023103818628763758 https://twitter.com/i/web/status/2023103818628763758"
X Link 2026-02-15T18:35Z 284.3K followers, [--] engagements

"@_apbright @SnazzyLabs @SawyerMerritt @Tesla talking about Apple Music you'd pay $10 a month for a feature on a phone you own outright do you not realize how insane that sounds"
X Link 2026-02-15T22:57Z 284.3K followers, [---] engagements

"You buy a license you never own it Would you say the same to Adobe Slack Microsoft with Office If you dont like it there are other options obviously. Rivian provides another option and product. I dont have to agree or like it to just accept it is what it is. And to be clear I was given an option and notice that if I wanted to own a perpetual license for a flat rate I could purchase it until yesterday. Again new owners are in a different boat. https://twitter.com/i/web/status/2023175563783917570 https://twitter.com/i/web/status/2023175563783917570"
X Link 2026-02-15T23:20Z 284.3K followers, [--] engagements

"@_apbright @DavidPiccinini2 @SnazzyLabs @SawyerMerritt @Tesla Are you illiterate"
X Link 2026-02-15T23:21Z 284.3K followers, [--] engagements

"You own a version of Microsoft office locked to a certain point. I too own Microsoft office [----] it has not received a single update since [----]. And if majority of people pirated adobe software adobe wouldnt be a many billion dollar company. SaaS is software as a service. Thats FSD. You still need to buy the phone or computer you access slack on. https://twitter.com/i/web/status/2023176617002995924 https://twitter.com/i/web/status/2023176617002995924"
X Link 2026-02-15T23:24Z 284.3K followers, [--] engagements

"@_apbright @DavidPiccinini2 @SnazzyLabs @SawyerMerritt @Tesla I didnt I chose to pay monthly once the product was good because I prefer to pay $100/m rather than $8000"
X Link 2026-02-15T23:25Z 284.3K followers, [--] engagements

"@_apbright @DavidPiccinini2 @SnazzyLabs @SawyerMerritt @Tesla Maybe it doesnt maybe it doesnt"
X Link 2026-02-15T23:30Z 284.3K followers, [--] engagements

"I have used my switch for maybe [--] hours since I bought it in 2018"
X Link 2026-02-15T23:56Z 284.3K followers, 14.9K engagements

"Raspberry Pi [--] in a box with a leaner Linux distro with the claw preinstalled on a 3D printed claw case If OpenAI were smart theyd make devices specifically to run OpenClaw on for cheaper than a Mac Mini. Apple $AAPL is stealing alot of the revenues right now via their Mac Minis. If OpenAI were smart theyd make devices specifically to run OpenClaw on for cheaper than a Mac Mini. Apple $AAPL is stealing alot of the revenues right now via their Mac Minis"
X Link 2026-02-16T00:52Z 284.3K followers, 15.5K engagements

"@Guille9215 Im just not a controller guy. I follow all the handhelds but cant get used to it. Have a Legion Go S and all"
X Link 2026-02-16T00:52Z 284.3K followers, [---] engagements

"@seempaq Im 100% serious"
X Link 2026-02-16T01:15Z 284.3K followers, [--] engagements

"RT @BenBajarin: Hard to really say past [----] for certainty at this point but its very easy to lay out the scenarios that keep memory const"
X Link 2026-02-16T15:54Z 284.3K followers, [--] engagements

"@tinygrad USB C portable ML accelerator"
X Link 2026-02-16T15:56Z 284.3K followers, [----] engagements

"Every mouse needs to be haptic This mouse is so sick https://t.co/KSfxCTIwXa This mouse is so sick https://t.co/KSfxCTIwXa"
X Link 2026-02-16T15:58Z 284.3K followers, 10.9K engagements

"@ajambrosino Honestly electron is probably better for this I like your UI/UX a lot All I want is to connect my iPhone to it lol"
X Link 2026-02-16T16:22Z 284.3K followers, [---] engagements

"@itsnoahd In terms of what specifically"
X Link 2026-02-16T16:38Z 284.3K followers, [---] engagements

"@Ben_jam_in__ Vibes are off and lacks nuance and detail in places it should have it"
X Link 2026-02-17T16:15Z 284.3K followers, [--] engagements

"@LukeWat48544961 @LuftkoppTim It was using the display like a person would. Scrolling the web working in photo editing apps word docs excel etc"
X Link 2026-02-18T15:24Z 284.3K followers, [--] engagements

"I need to remember to stop enabling OneDrive on computers I'll play games on It keeps syncing Overwatch settings across 2-3 laptops with completely different specs and needed settings and it's driving me insane"
X Link 2026-02-11T03:30Z 284.3K followers, [----] engagements

"Working on a web battery loop and noticed that edge was using more power than Chrome so I had Codex run the same loop with Edge vs. Google Chrome Edge used 7W more power for the exact same loop I'd recommend not using Edge as your browser on Windows if you value battery"
X Link 2026-02-16T18:07Z 284.3K followers, 10.5K engagements

"@scaling01 The idea youre fitting 1T parameter model on a [--] consumer RTX cards is absurd Sure if your host system has 2TB of memory and you do hybrid CPU inference but enjoy your [--] token per second and 4K context window lmfao"
X Link 2026-02-16T21:41Z 284.3K followers, [----] engagements

"Running battery tests at [---] nits is stupid"
X Link 2026-02-17T14:34Z 284.3K followers, 12.8K engagements

"This is a good post but a few things Id point out and note. I love hosting and running my own models I use GLM-5 on a Mac Studio cluster and have a DGX Spark always cranking tokens out that being said [--]. It does not make sense for any single person to buy $20K worth of Mac Studios for AI inference. If youre at an office and [--] or more people can use those Mac Studios as computers that have the model running as well thats a good use. Buying them specifically for AI is a waste of money. Nobody should be recommending this. Its an outstanding setup if you can afford it by all means But these are"
X Link 2026-02-18T00:00Z 284.3K followers, [----] engagements

"If Apple and Google (via Android products powered by Google services) have AI hardware that does the same things as Meta OpenAI or others to be created Would you rather stay with what you know or try the others Assume near feature parity or [--] month lag time to new feature Apple/Google (Android) Others Apple/Google (Android) Others"
X Link 2026-02-18T00:09Z 284.3K followers, [----] engagements

"@LuftkoppTim Really only would be the memory. Die size is roughly equal and N3E and N3P are the same per wafer Might be easier to have [--] SKU on [--] node vs [--] on [--] nodes"
X Link 2026-02-18T23:44Z 284.3K followers, [---] engagements

"Damn I had Claude Cowork explain it's harness to me tools and prompt and all then create documentation on how to recreate itself as model agnostic using ai sdk then I handed those [----] lines of documentation to GPT [---] Codex and it was able to one-shot a working version"
X Link 2026-02-08T03:39Z 284.3K followers, 34.3K engagements

"Had GPT [---] Codex make a native app for Google Messages on macOS Few things to fix up but can send/receive messages (including RCS) on my Pixel from my Mac"
X Link 2026-02-08T04:44Z 284.3K followers, 115.1K engagements

"GPT-5.3-Spark is pretty nice for the type of stuff youd imagine you want a bunch of quick changes to like tweaking UI/copy/design stuff or researching a codebase for info Only thing Id really want is a larger context window which seems to be coming later"
X Link 2026-02-12T22:03Z 284.3K followers, [----] engagements

"@ljnobre @LuftkoppTim Its the same chip as the pro iPhone with a core disabled its binning not a different design Maybe they will who knows"
X Link 2026-02-18T23:59Z 284.3K followers, [---] engagements

"I do not like this tweet. It reads as though something is wrong with the other panther lake SKUs. There is not. It's just a different SKU of the chip with less CPU cores. The end products it comes in maybe more than Lunar Lake but that's because Lunar Lake systems are a generation old and discounted. The thing is Panther Lake systems as a whole are designed way better than Lunar Lake ones. Most are more premium and simply better products. Nobody really buys laptops for the chip outside of dGPU versions they choose a laptop and select SoC based on price/availability in that chassis. The"
X Link 2026-02-19T00:33Z 284.3K followers, [----] engagements

"This is mine over [--] ChatGPT Pro subs it is worth noting tho that at least for Codex tbd on CC the inference stack is different and slower. Likely higher batch size so more optimized for Codex usage lower cost to run more users. And inference is profitable just upfront training and R&D is EXPENSIVE https://twitter.com/i/web/status/2022822343505023004 https://twitter.com/i/web/status/2022822343505023004"
X Link 2026-02-14T23:56Z 284.3K followers, [----] engagements

"AI playlist creation in Apple Music"
X Link 2026-02-16T20:16Z 284.3K followers, 111.7K engagements

"Panther Lake + Overwatch update really has made me way more interested in playing games again lol Poor Ramattra"
X Link 2026-02-17T00:18Z 284.3K followers, [----] engagements

"This runs on device using the Apple Foundation Model Im pretty sure. AI playlist creation in Apple Music https://t.co/dteWoKRK1b AI playlist creation in Apple Music https://t.co/dteWoKRK1b"
X Link 2026-02-17T01:05Z 284.3K followers, 65.6K engagements

"@bxlewi1 Everyone does every battery stat you see from any test is at [---] nits All of them They can't even see the content that's running* depends on display panel cover material coating texture panel type and like a million other things I hate it"
X Link 2026-02-17T15:45Z 284.3K followers, [---] engagements

"Jr Analysts in Midtown just let out a screech in unison For Claude in Excel users our add-in now supports MCP connectors letting Claude work with tools like S&P Global LSEG Daloopa PitchBook Moodys and FactSet. Pull in context from outside your spreadsheet without ever leaving Excel. https://t.co/hrkKpb2ePw For Claude in Excel users our add-in now supports MCP connectors letting Claude work with tools like S&P Global LSEG Daloopa PitchBook Moodys and FactSet. Pull in context from outside your spreadsheet without ever leaving Excel. https://t.co/hrkKpb2ePw"
X Link 2026-02-17T18:22Z 284.3K followers, 430.1K engagements

"Just registered for Google IO Whos going this year I am REALLY excited for this one"
X Link 2026-02-17T20:55Z 284.3K followers, [----] engagements

"Overwatch keeps putting me in matches with people who have like 4-6k hours into the game and level [----] or or some stuff and I need to just sweat a little Is fun tho"
X Link 2026-02-18T03:12Z 284.3K followers, [----] engagements

"My review of the RTX [----] last year was pretty simple it's the best gaming GPU ever but we need AI upscaling to see performance gains. Raster arch is pretty mature. Adding tensor cores and AI models is the only cost effective way. NVIDIA DLSS [---] Becomes The Top Choice In Blind Test Vs Native Res & AMD FSR Upscaling https://t.co/DplA5Os7RY https://t.co/DplA5Os7RY NVIDIA DLSS [---] Becomes The Top Choice In Blind Test Vs Native Res & AMD FSR Upscaling https://t.co/DplA5Os7RY https://t.co/DplA5Os7RY"
X Link 2026-02-18T17:03Z 284.3K followers, [----] engagements

"it so feels like they want codex to be the default harness for chatgpt but cloud sandboxes and a better general prompt all your files (cloud storage) being managed by an agent and subagents that can chat with you or delegate tasks and work with your files and connectors"
X Link 2026-02-18T17:07Z 284.3K followers, [----] engagements

"Website that buys used iPhone's says they are buying a lot of the newest iPhone. What a poor report. I was hoping the report would say what they are trade in for how this compares to previous generations anything else. Nope. iPhone [--] Pro Max Curiously Becomes Most Traded-In Smartphone https://t.co/dCRNmYXqqI https://t.co/2WroOP5Zk6 iPhone [--] Pro Max Curiously Becomes Most Traded-In Smartphone https://t.co/dCRNmYXqqI https://t.co/2WroOP5Zk6"
X Link 2026-02-18T17:43Z 284.3K followers, 12K engagements

"@Solaawodiya I'd compare it to MacBook Air and it compares super well"
X Link 2026-02-18T21:55Z 284.3K followers, [--] engagements

"@PatrickMoorhead I sont think so since apparently MacBook Air will get M5 plus A series uses infopop packaging with memory on package so itll be [--] memory config Then ofc [--] CPU/GPU cores vs [--] for M5 This is universal for A series so its all which gen IP/memory they want"
X Link 2026-02-18T23:43Z 284.3K followers, [---] engagements

"@XPh03n @PatrickMoorhead My guess is they offer a [---] [---] 1tb option and color and that's it 4-5 colors 2-3 storage options anything else you might as well get a MacBook Air"
X Link 2026-02-19T01:39Z 284.3K followers, [---] engagements

"This is actually pretty cool from Gemini"
X Link 2026-01-22T17:16Z 284.3K followers, 328.4K engagements

"So Codex has memory now you can enable it by adding the following to your config.toml features memory_tool = true sqlite = true but be warned it used 50% of my [--] hour limit in like [--] minutes lmao. BURNS tokens but seems to help"
X Link 2026-02-12T00:42Z 284.3K followers, 49K engagements

"I shit you not this was [--] minutes of Codex memory running 25% of my [--] hour token allotment tokens in [--] minutes It's worth it for memory but holy crap"
X Link 2026-02-12T00:50Z 284.3K followers, 17.9K engagements

"This mouse is so sick https://x.com/mweinbach/status/2023417028523209203/photo/1 https://x.com/mweinbach/status/2023417028523209203/photo/1"
X Link 2026-02-16T15:19Z 284.3K followers, 17.1K engagements

"Thinking about the cheap MacBook Imagine they make a Mac Nano that uses the same A-series chip in N even smaller chassis than the Mac Mini with just USB C like [--] ports including charging and display out. Maybe something the size of an iPhone but its a Mac for $299-399"
X Link 2026-02-16T16:04Z 284.3K followers, 24.1K engagements

"Ramattras ultimate line being suffer as I have is pretty brutal given how powerful his ultimate is"
X Link 2026-02-17T02:49Z 284.3K followers, [----] engagements

"I am having too much fun"
X Link 2026-02-17T02:52Z 284.3K followers, [----] engagements

"So Grok [----] just sends [--] model instances to answer your query and answer the question I like this approach but hard to vibe test the model. Running the same query a few times is usually the best way to get a higher quality response this native experience is nice"
X Link 2026-02-17T13:54Z 284.3K followers, [----] engagements

"Charging up all my Snapdragon X Elite laptops. I need to be better at keeping these things updated lol"
X Link 2026-02-18T21:00Z 284.3K followers, [----] engagements

"@SIGKITTEN i mean yea but like use ai sdk and agents sdk connector and build on open framework and use cheap tokens"
X Link 2026-02-18T21:55Z 284.3K followers, [--] engagements

"I mean. Samsung did make the best Windows laptop Ive personally used And its very MacBook-y. I think thats a really good thing"
X Link 2026-02-05T23:59Z 284.3K followers, 644.6K engagements

"Whoop membership is now included with Chase Sapphire This is huge for chase sapphire card holders because whoop is actually great @mweinbach this is a big deal https://t.co/Z0EkX3icIP @mweinbach this is a big deal https://t.co/Z0EkX3icIP"
X Link 2026-02-12T19:28Z 284.3K followers, 474.2K engagements

"I really hope its M5 Max I want it just so I can run local models without that solid 10-45 second wait for prefill to process It unfortunately ruins the experience of the great local models if you are doing anything real time. For async who cares Apple is holding a product launch in New York on March [--]. https://t.co/j0hYQ9sfqd Apple is holding a product launch in New York on March [--]. https://t.co/j0hYQ9sfqd"
X Link 2026-02-16T16:16Z 284.3K followers, 17.3K engagements

"Literally none of the model assumptions here around Claude DeepSeek and GPT [---] make sense or are even possible Burry got caught in the slop hopefully intelligent investors weed this info out Convergence commoditzation compression and reflexivity - the [--] horsemen of the data center buildout apocalypse. DeepSeek V4 launches mid-February [----] 1T parameters 1M token context [--] architectural innovations @ 10-40x lower than Western Comps $NVDA https://t.co/78IwAQx6yz Convergence commoditzation compression and reflexivity - the [--] horsemen of the data center buildout apocalypse. DeepSeek V4 launches"
X Link 2026-02-16T23:05Z 284.3K followers, [----] engagements

"RT @AnushElangovan: BS. Right on target for H2 2026"
X Link 2026-02-17T17:59Z 284.3K followers, [--] engagements

"I still think the cheap MacBook is A19 Pro A18 Pro doesn't make sense when the A19 Pro CPU/GPU IP and memory makes way more sense for a MacBook Maybe they'll do it because they made too many which I also heard but from a product standpoint A19 Pro makes too much sense"
X Link 2026-02-18T22:11Z 284.3K followers, 16.6K engagements

"@Prakhar6200 @ExoticSpice101 Most brands have only launched their ultra premium and premium models so far More soon"
X Link 2026-02-19T02:22Z 284.3K followers, [--] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing