Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

timelapse #85 (27.5 hrs): - currently cant rely on any other coding models except grok code fast 1 + grok 4 fast (for complex reasoning grok 4 fast is 20 cents for 1M tokens) - wrote qwen3-next trainer entirely from scratch to make it more managable - each piece...

elliotarledge's profile picture

Elliot Arledge

32,020 subscribers

283,820 Aufrufe • vor 10 Monaten •via X (Twitter)

BildungWissenschaft & Technologie
LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

timelapse #86 (15 hrs): - got my first OOM on 8xB200 node - defaulting back to grok-code-fast-1, the fastest reliable coding model with by far most intuitive instruction following, combined with grok 4 fast reasoning to plan before i let grok code work its magic - drank 2 large tim hortons iced capps, loaded myself w/ creatine, daily nootropics - tried out gpt-5-codex but it simply doesnt match the speed i require when i go deep into one thing at a time sequentially - got caught watching youtube videos in the middle, need to make sure i block any and all content that could get in my way - caught up on all book revisions so getting super ahead with other chapters - developed an overnight addiction to switching color themes in cursor - did some pair programming w/ Kearm h/eng using Tuple on free trial - applying for O-1
0:27

timelapse #86 (15 hrs): - got my first OOM on 8xB200 node - defaulting back to grok-code-fast-1, the fastest reliable coding model with by far most intuitive instruction following, combined with grok 4 fast reasoning to plan before i let grok code work its magic - drank 2 large tim hortons iced capps, loaded myself w/ creatine, daily nootropics - tried out gpt-5-codex but it simply doesnt match the speed i require when i go deep into one thing at a time sequentially - got caught watching youtube videos in the middle, need to make sure i block any and all content that could get in my way - caught up on all book revisions so getting super ahead with other chapters - developed an overnight addiction to switching color themes in cursor - did some pair programming w/ Kearm h/eng using Tuple on free trial - applying for O-1

Elliot Arledge

102,959 Aufrufe • vor 10 Monaten

timelapse #83 (22 hrs): - it was very easy to dive super deep into anything i needed to (this is what i focused on today because not all days are like this) - finding the grok code fast 1 + grok 4 for deep thinking and verification combo to be super useful in cursor. speed was solid - hard to imagine myself spending many more mental clock cycles in a 24 hr period - had to pull out qwen3-next’s gated deltanet + linear attention from bleeding edge hf transformers to begin implementing a multi-gpu fp8 trainer from scratch. this is so damn bleeding edge and i underestimated how much effort this has and will require lol - lots of diet coke and oats - shipped the template which the core chapters of my book will be built on: - all im missing now is flash attention 1/2 mastery (fa2 tmrw), intuition on making topk faster (for arbitrary row length), what i should and shouldnt teach in cutlass/cute, hopper/blackwell gemm kernel mastery (down to fp4) — shoutout to Pranjal for making this easier for me. his blog post is amazing - caught up w/ Mati Roy - im feeling great mentally but not so good physically as im writing this and about to pass out
0:30

timelapse #83 (22 hrs): - it was very easy to dive super deep into anything i needed to (this is what i focused on today because not all days are like this) - finding the grok code fast 1 + grok 4 for deep thinking and verification combo to be super useful in cursor. speed was solid - hard to imagine myself spending many more mental clock cycles in a 24 hr period - had to pull out qwen3-next’s gated deltanet + linear attention from bleeding edge hf transformers to begin implementing a multi-gpu fp8 trainer from scratch. this is so damn bleeding edge and i underestimated how much effort this has and will require lol - lots of diet coke and oats - shipped the template which the core chapters of my book will be built on: - all im missing now is flash attention 1/2 mastery (fa2 tmrw), intuition on making topk faster (for arbitrary row length), what i should and shouldnt teach in cutlass/cute, hopper/blackwell gemm kernel mastery (down to fp4) — shoutout to Pranjal for making this easier for me. his blog post is amazing - caught up w/ Mati Roy - im feeling great mentally but not so good physically as im writing this and about to pass out

Elliot Arledge

2,255,154 Aufrufe • vor 10 Monaten

GROK CODE FAST 1 TAKES #1 ON OPENROUTER, DEVELOPERS PICK REAL PERFORMANCE OVER HYPE Grok Code Fast 1 just locked in the top spot on OpenRouter’s December 2025 monthly leaderboard, and this one actually matters. Real workloads. Real builders choosing what they trust when things break, scale, or need to ship. According to OpenRouter data, Grok Code Fast 1 processed billions of tokens in production, clearing every major competitor. Claude Sonnet 4, GPT-5 variants, Gemini, DeepSeek , all solid models, all left behind. Developers didn’t debate it on X. They just routed traffic. The reasons are practical, not philosophical. Grok Code Fast 1 is fast. Like actually fast. Low latency, massive 2M context window, strong tool calling, and code outputs that don’t collapse the moment you introduce edge cases. What keeps coming up from builders is reliability. Fewer hallucinations. Better memory across long sessions. It doesn’t fight you. It doesn’t nanny you. It just does the work. Yes, the free trial helped drive initial adoption. But free doesn’t explain retention at this scale. People stayed because it worked better for serious tasks, not toy prompts. The bigger signal is this: when you strip away marketing, guardrail theater, and benchmark gaming, developers reward models that prioritize truth, speed, and usefulness. Source: X Freeze, OpenRouter, Grok
0:30

GROK CODE FAST 1 TAKES #1 ON OPENROUTER, DEVELOPERS PICK REAL PERFORMANCE OVER HYPE Grok Code Fast 1 just locked in the top spot on OpenRouter’s December 2025 monthly leaderboard, and this one actually matters. Real workloads. Real builders choosing what they trust when things break, scale, or need to ship. According to OpenRouter data, Grok Code Fast 1 processed billions of tokens in production, clearing every major competitor. Claude Sonnet 4, GPT-5 variants, Gemini, DeepSeek , all solid models, all left behind. Developers didn’t debate it on X. They just routed traffic. The reasons are practical, not philosophical. Grok Code Fast 1 is fast. Like actually fast. Low latency, massive 2M context window, strong tool calling, and code outputs that don’t collapse the moment you introduce edge cases. What keeps coming up from builders is reliability. Fewer hallucinations. Better memory across long sessions. It doesn’t fight you. It doesn’t nanny you. It just does the work. Yes, the free trial helped drive initial adoption. But free doesn’t explain retention at this scale. People stayed because it worked better for serious tasks, not toy prompts. The bigger signal is this: when you strip away marketing, guardrail theater, and benchmark gaming, developers reward models that prioritize truth, speed, and usefulness. Source: X Freeze, OpenRouter, Grok

Mario Nawfal

39,820 Aufrufe • vor 7 Monaten

You can launch Grok on Windows with just one click - no need to open Chrome or Edge every time It runs as a lightweight app using the PWA method, so it feels fast and native Here’s how to set it up: Step 1 → Open in Chrome or Edge Step 2 → Click the little “install app” icon at the top-right of the address bar and hit “Install” Step 3 → Right‑click the new Grok icon in your taskbar and click “Pin to taskbar” Now Grok opens instantly without loading the full browser, which makes it much snappier and way more efficient You can also assign a hotkey to the shortcut (e.g., Ctrl + Alt + G) so Grok is always one key combo away
0:24

You can launch Grok on Windows with just one click - no need to open Chrome or Edge every time It runs as a lightweight app using the PWA method, so it feels fast and native Here’s how to set it up: Step 1 → Open in Chrome or Edge Step 2 → Click the little “install app” icon at the top-right of the address bar and hit “Install” Step 3 → Right‑click the new Grok icon in your taskbar and click “Pin to taskbar” Now Grok opens instantly without loading the full browser, which makes it much snappier and way more efficient You can also assign a hotkey to the shortcut (e.g., Ctrl + Alt + G) so Grok is always one key combo away

X Freeze

27,670 Aufrufe • vor 7 Monaten

In the latest Box AI Enterprise Evaluation, we tested xAI's Grok 4. We have seen how this new model moves beyond surface-level retrieval to tackle sophisticated business logic, perform precise calculations, make inferences based on qualitative patterns, and pinpoint critical contract clauses. 👉 Key Highlights: ↳ When analyzing company financial data, Grok 4 correctly performed multi-step tasks, followed sequential logic, and made accurate calculations to determine gross margins and performance comparisons. ↳ In reviewing information from text passages, Grok 4 showed advanced qualitative reasoning, comparing stylistic elements like tone, perspective, and vocabulary to correctly group passages and identify the number of authors. ↳ It excelled in extracting complex information from contracts, identifying detailed clauses like uncapped liability and profit sharing, as well as analyzing the implications of interdependent terms in agreements. 💡 Why It Matters: ↳ For Legal and Finance teams, Grok 4’s improved ability to handle calculations and interpret complex clauses makes it a powerful tool for in-depth contract review and financial analysis. ↳ For researchers, the model's advanced analytical capabilities can help deconstruct and synthesize information from dense technical papers. 👉 The Takeaway Overall, Grok 4 shows measurable advancement in sequential logic, numerical precision, and domain-specific language understanding. The model’s ability to blend quantitative and qualitative reasoning widens the range of workflows that can be automated inside Box.
2:29

In the latest Box AI Enterprise Evaluation, we tested xAI's Grok 4. We have seen how this new model moves beyond surface-level retrieval to tackle sophisticated business logic, perform precise calculations, make inferences based on qualitative patterns, and pinpoint critical contract clauses. 👉 Key Highlights: ↳ When analyzing company financial data, Grok 4 correctly performed multi-step tasks, followed sequential logic, and made accurate calculations to determine gross margins and performance comparisons. ↳ In reviewing information from text passages, Grok 4 showed advanced qualitative reasoning, comparing stylistic elements like tone, perspective, and vocabulary to correctly group passages and identify the number of authors. ↳ It excelled in extracting complex information from contracts, identifying detailed clauses like uncapped liability and profit sharing, as well as analyzing the implications of interdependent terms in agreements. 💡 Why It Matters: ↳ For Legal and Finance teams, Grok 4’s improved ability to handle calculations and interpret complex clauses makes it a powerful tool for in-depth contract review and financial analysis. ↳ For researchers, the model's advanced analytical capabilities can help deconstruct and synthesize information from dense technical papers. 👉 The Takeaway Overall, Grok 4 shows measurable advancement in sequential logic, numerical precision, and domain-specific language understanding. The model’s ability to blend quantitative and qualitative reasoning widens the range of workflows that can be automated inside Box.

Box

1,973,403 Aufrufe • vor 1 Jahr

Inspired by Grok as a developer and a heavy gamer for over 15 years, I spent some time last week building a few things. Thrilled to unveil my latest creation: an infinite runner game built almost entirely by Grok from xAI! This project showcases the incredible power of AI in game development. Grok handled everything—from designing the game mechanics to writing the code and even helping me debug issues along the way. I brought it to life using some amazing free assets from a treasure trove for indie developers. You can play the game now at Elon Musk, I’d be honored if you checked it out. AI is revolutionizing game development, and Grok is at the forefront with its outstanding capabilities. It’s more than a tool—it’s like a tireless co-developer. Grok grasps complex concepts, provides suggestions, and turns rough ideas into working code fast. For this infinite runner, it crafted smooth player controls, randomized obstacle generation, and an engaging scoring system, letting me focus on the overall vision. And cross_protocol, founded by Henry @CROSS is set to harness AI’s full potential in gaming, pushing the boundaries even further. This is just the beginning. With Grok’s help, I’m planning future projects: 1) Physics-based puzzle game where players tweak gravity and momentum to solve puzzles 2) 2D RPG with deep storytelling and branching dialogue 3) Fast-paced 3D shooter with immersive worlds Each genre requires unique skills, but Grok’s versatility makes it ideal for all of them. It adapts to any challenge—be it physics simulations, character AI, or level design—producing results that could rival a full dev team. AI like Grok is opening up creative doors I couldn’t tackle alone, and I can’t wait to see what’s next. Stay tuned for more.
0:20

Inspired by Grok as a developer and a heavy gamer for over 15 years, I spent some time last week building a few things. Thrilled to unveil my latest creation: an infinite runner game built almost entirely by Grok from xAI! This project showcases the incredible power of AI in game development. Grok handled everything—from designing the game mechanics to writing the code and even helping me debug issues along the way. I brought it to life using some amazing free assets from a treasure trove for indie developers. You can play the game now at Elon Musk, I’d be honored if you checked it out. AI is revolutionizing game development, and Grok is at the forefront with its outstanding capabilities. It’s more than a tool—it’s like a tireless co-developer. Grok grasps complex concepts, provides suggestions, and turns rough ideas into working code fast. For this infinite runner, it crafted smooth player controls, randomized obstacle generation, and an engaging scoring system, letting me focus on the overall vision. And cross_protocol, founded by Henry @CROSS is set to harness AI’s full potential in gaming, pushing the boundaries even further. This is just the beginning. With Grok’s help, I’m planning future projects: 1) Physics-based puzzle game where players tweak gravity and momentum to solve puzzles 2) 2D RPG with deep storytelling and branching dialogue 3) Fast-paced 3D shooter with immersive worlds Each genre requires unique skills, but Grok’s versatility makes it ideal for all of them. It adapts to any challenge—be it physics simulations, character AI, or level design—producing results that could rival a full dev team. AI like Grok is opening up creative doors I couldn’t tackle alone, and I can’t wait to see what’s next. Stay tuned for more.

J

99,726 Aufrufe • vor 1 Jahr

“ xAI team is currently working heavily on coding models. Right now, the main focus is training a specialized coding model that will be both fast and smart. I believe we’ll share it with you guys in a few weeks. That’s exciting. Second ,after coding, we all see that the main weakness of Grok 4 is its multimodal capabilities. In fact, it was so bad that Grok was effectively looking at the world while squinting through glass trying to see blurry features and make sense of them. The most immediate improvement we’re stepping on with the next generation pre trained model is huge gains in image understanding, video understanding, and audio. Right now, the model can hear and see the world just like any of you. And with all the tools and other agents it can talk to, we’re going to see a huge unlock for many different application layers once multimodal agents arrive.” — xAI Team
2:01

“ xAI team is currently working heavily on coding models. Right now, the main focus is training a specialized coding model that will be both fast and smart. I believe we’ll share it with you guys in a few weeks. That’s exciting. Second ,after coding, we all see that the main weakness of Grok 4 is its multimodal capabilities. In fact, it was so bad that Grok was effectively looking at the world while squinting through glass trying to see blurry features and make sense of them. The most immediate improvement we’re stepping on with the next generation pre trained model is huge gains in image understanding, video understanding, and audio. Right now, the model can hear and see the world just like any of you. And with all the tools and other agents it can talk to, we’re going to see a huge unlock for many different application layers once multimodal agents arrive.” — xAI Team

Apurv Kochara

88,142 Aufrufe • vor 5 Monaten

"At the Edge of the Galaxy" – Episode 1 I created this video entirely using AI, specifically the Grok Imagine software. I developed the script with Grok . Then I produced it using all the tools available in Grok Grok Imagine , particularly the Agent mode, which saves a tremendous amount of time. I made this video alone, using just a standard computer. I was blown away by the ability to create dialogue that sounds authentic and worlds filled with such lifelike creatures. The consistency of the characters and locations is excellent, and the realism is breathtaking. I don't see what is stopping anyone today from creating a film entirely conceived using AI with Grok Imagine. If you have any questions, please feel free to ask. I will be happy to answer them. If you would like to support my work, the best way to help is to repost the video. Thank you in advance. As it happens, the video is 4 minutes and 20 seconds long (honestly, I didn't do that on purpose)—that ought to bring me good luck, right? Elon Musk
4:20

"At the Edge of the Galaxy" – Episode 1 I created this video entirely using AI, specifically the Grok Imagine software. I developed the script with Grok . Then I produced it using all the tools available in Grok Grok Imagine , particularly the Agent mode, which saves a tremendous amount of time. I made this video alone, using just a standard computer. I was blown away by the ability to create dialogue that sounds authentic and worlds filled with such lifelike creatures. The consistency of the characters and locations is excellent, and the realism is breathtaking. I don't see what is stopping anyone today from creating a film entirely conceived using AI with Grok Imagine. If you have any questions, please feel free to ask. I will be happy to answer them. If you would like to support my work, the best way to help is to repost the video. Thank you in advance. As it happens, the video is 4 minutes and 20 seconds long (honestly, I didn't do that on purpose)—that ought to bring me good luck, right? Elon Musk

Déborah

117,467 Aufrufe • vor 23 Tagen

Which LLM reasons best when it doesn't have all the information? Enter LLM Poker Arena to find out. It's a Poker Playing benchmark where top reasoning models play Texas Hold'em poker against each other. Claude Opus 4.5, GPT-5.2, Gemini 2.5 Pro, and Grok 4 all sit at the same table and play full tournaments to see who finishes with the chips. Poker is very different when it comes to reasoning. It has to balance probabilistic reasoning, opponent modeling and make decisions under uncertainty. Poker is an interesting evaluation because it tests reasoning under incomplete information, something most coding benchmarks do not capture. In this tournaments the rules are: - Each LLM starts with $1,000 chips - Small and big blinds start at $25 / $50 - Blinds double every 3 minutes - All models run in their reasoning or thinking modes After the first 5 tournaments: - Claude Opus 4.5 with Thinking has 3 wins - GPT-5.2 has 2 wins - Grok 4 and Gemini 2.5 Pro have 0 wins Early results suggest Claude performs quite well at poker as well. Also five is a very small sample size. Planning to run many more tournaments, publish the full benchmark data and add a prediction market on top of it. Thanks for the suggestion clipz. Much more coming as part of Poker Cities !! This was built on Replit ⠕ using their AI integrations, which made it straightforward to connect Claude, GPT, and Gemini. What model do you think wins after 100 tournaments?
0:22

Which LLM reasons best when it doesn't have all the information? Enter LLM Poker Arena to find out. It's a Poker Playing benchmark where top reasoning models play Texas Hold'em poker against each other. Claude Opus 4.5, GPT-5.2, Gemini 2.5 Pro, and Grok 4 all sit at the same table and play full tournaments to see who finishes with the chips. Poker is very different when it comes to reasoning. It has to balance probabilistic reasoning, opponent modeling and make decisions under uncertainty. Poker is an interesting evaluation because it tests reasoning under incomplete information, something most coding benchmarks do not capture. In this tournaments the rules are: - Each LLM starts with $1,000 chips - Small and big blinds start at $25 / $50 - Blinds double every 3 minutes - All models run in their reasoning or thinking modes After the first 5 tournaments: - Claude Opus 4.5 with Thinking has 3 wins - GPT-5.2 has 2 wins - Grok 4 and Gemini 2.5 Pro have 0 wins Early results suggest Claude performs quite well at poker as well. Also five is a very small sample size. Planning to run many more tournaments, publish the full benchmark data and add a prediction market on top of it. Thanks for the suggestion clipz. Much more coming as part of Poker Cities !! This was built on Replit ⠕ using their AI integrations, which made it straightforward to connect Claude, GPT, and Gemini. What model do you think wins after 100 tournaments?

Anshul Dhawan

32,192 Aufrufe • vor 6 Monaten

‼️🇺🇸 NEW: Using Grok 4, Occam's razor, first-principle reasoning, and primary evidence in our X Space, we were able to isolate the suspicious ANOMALIES surrounding the Epstein-Maxwell Case: 1: No plea deal offered by the government - VERY unusual for the government not to offer ANY kind of deal in exchange for information about accomplices 👀 2: Jury instructed in such a way that it left room for conviction without having to implicate 3rd parties for certain s3x acts 🤨 - Grok then goes on to list out ALL of the inconsistencies throughout the Epstein case that highlight the presence of what seems to be an elite transnational network with damning evidence on each other, with its influence being wielded in the Trump administration TODAY. 👀 Even the legacy AND new media seem to be already infiltrated by the network. This network must be STOPPED. At ALL costs!
16:53

‼️🇺🇸 NEW: Using Grok 4, Occam's razor, first-principle reasoning, and primary evidence in our X Space, we were able to isolate the suspicious ANOMALIES surrounding the Epstein-Maxwell Case: 1: No plea deal offered by the government - VERY unusual for the government not to offer ANY kind of deal in exchange for information about accomplices 👀 2: Jury instructed in such a way that it left room for conviction without having to implicate 3rd parties for certain s3x acts 🤨 - Grok then goes on to list out ALL of the inconsistencies throughout the Epstein case that highlight the presence of what seems to be an elite transnational network with damning evidence on each other, with its influence being wielded in the Trump administration TODAY. 👀 Even the legacy AND new media seem to be already infiltrated by the network. This network must be STOPPED. At ALL costs!

Diligent Denizen 🇺🇸

46,249 Aufrufe • vor 1 Jahr

43 Years Ago Today — On June 3, 1983, WarGames hit theaters and gave us one of the most chilling tech thrillers ever made. A curious kid hacks into a military supercomputer, starts a game of “Global Thermonuclear War,” and nearly triggers the real thing. Matthew Broderick, Ally Sheedy, and that unforgettable countdown… pure 80s perfection. Each year I ask Grok how long it would take for it to simulate the same scenario, last year it was about 10 minutes...with a special conditon, Grok emphasised each year... see below. This year, 2026 on the anniversary of the film, I asked Grok (xAI’s AI) the big question from the film: How long would it take a modern AI like Grok to simulate a full Global Thermonuclear War scenario? Answer: Seconds for high-level insights, minutes for detailed probabilistic simulations — a massive leap from WOPR’s 26+ hours in 1983. Computing power and AI have come that far! Each year Grok after my asking this, Grok concludes with... "Yet the movie’s ultimate lesson remains as true today as it was then: “The only winning move is not to play.”" At least we know Grok won't be starting WWIII... I have seen other Ai's coming out with chilling answers, which don't end well for us... us being humans. Wargames the movie is a rollercoaster ride, that still holds up today. Who else still loves this film?
2:13

43 Years Ago Today — On June 3, 1983, WarGames hit theaters and gave us one of the most chilling tech thrillers ever made. A curious kid hacks into a military supercomputer, starts a game of “Global Thermonuclear War,” and nearly triggers the real thing. Matthew Broderick, Ally Sheedy, and that unforgettable countdown… pure 80s perfection. Each year I ask Grok how long it would take for it to simulate the same scenario, last year it was about 10 minutes...with a special conditon, Grok emphasised each year... see below. This year, 2026 on the anniversary of the film, I asked Grok (xAI’s AI) the big question from the film: How long would it take a modern AI like Grok to simulate a full Global Thermonuclear War scenario? Answer: Seconds for high-level insights, minutes for detailed probabilistic simulations — a massive leap from WOPR’s 26+ hours in 1983. Computing power and AI have come that far! Each year Grok after my asking this, Grok concludes with... "Yet the movie’s ultimate lesson remains as true today as it was then: “The only winning move is not to play.”" At least we know Grok won't be starting WWIII... I have seen other Ai's coming out with chilling answers, which don't end well for us... us being humans. Wargames the movie is a rollercoaster ride, that still holds up today. Who else still loves this film?

Commodore Computer Museum 🕹

16,892 Aufrufe • vor 1 Monat

This is something no other AI and no other car company can do right now, except Grok, fully integrated with FSD (Supervised) v14.2.25. Today, I gave it a wild, real-world task: • Drive to Apple HQ • Then Tesla Engineering HQ • Then a Warriors game at Chase Center SF • Then the beach • Then Asian dinner It got it all ready from the beginning. Mid-route I said I really needed to use the bathroom! It added it as the first stop. Then I added a Tesla Supercharger on the fly for after the Warriors game. It added it. AND stopped for grandparents and baby crossing the street. Then I asked for the weather and exactly how much I should $ budget for the entire day. No friction. No talking back. Just getting it done insanely fast. It just handled it all naturally, intelligently, all in real time. Better than any driver, for sure… This is reasoning + autonomy + AI + real life, all working beautifully in synch. No other car, AI, is close to this capability imo. This is a MAJOR update. I honestly feel if this was any other company, they would’ve had a whole event just to showcase this feature.
2:54

This is something no other AI and no other car company can do right now, except Grok, fully integrated with FSD (Supervised) v14.2.25. Today, I gave it a wild, real-world task: • Drive to Apple HQ • Then Tesla Engineering HQ • Then a Warriors game at Chase Center SF • Then the beach • Then Asian dinner It got it all ready from the beginning. Mid-route I said I really needed to use the bathroom! It added it as the first stop. Then I added a Tesla Supercharger on the fly for after the Warriors game. It added it. AND stopped for grandparents and baby crossing the street. Then I asked for the weather and exactly how much I should $ budget for the entire day. No friction. No talking back. Just getting it done insanely fast. It just handled it all naturally, intelligently, all in real time. Better than any driver, for sure… This is reasoning + autonomy + AI + real life, all working beautifully in synch. No other car, AI, is close to this capability imo. This is a MAJOR update. I honestly feel if this was any other company, they would’ve had a whole event just to showcase this feature.

Teslaconomics

258,175 Aufrufe • vor 7 Monaten

I cant believe this guy just made a permanent solution to context bloat and open sourced it all! when we tested this tool (Context+) for solving an issue on the OpenCode repository, the agent using this tool used ~6.5k fewer tokens, found the code and fixed it in half the time! the results were surprising: 6 to 10k tokens saved per prompt, completed task in ~2 minutes while the agent running without the tool took ~4 mins for the same and got stuck in loops bro built an entire beast by using all the modern tools that we could think of: undo trees, semantic search by meaning (by haskellforall), advanced refactoring, blast radius, advanced file context trees, restore points... i can keep going on semantic code search and context trees are the future of agentic coding and this tool proves it the feature i loved the most is semantic search and how it gets things done 2x faster with least possible tokens it makes an agent that actually knows what it’s doing and not just guessing, it makes meaning from your code similar to RAG. if you aren't optimizing your context, you are just burning money the developer says this tool is still under development, it can have unexpected behavior and the docs need updates but the video shows the reality of how fast it can be github: get here:
0:46

I cant believe this guy just made a permanent solution to context bloat and open sourced it all! when we tested this tool (Context+) for solving an issue on the OpenCode repository, the agent using this tool used ~6.5k fewer tokens, found the code and fixed it in half the time! the results were surprising: 6 to 10k tokens saved per prompt, completed task in ~2 minutes while the agent running without the tool took ~4 mins for the same and got stuck in loops bro built an entire beast by using all the modern tools that we could think of: undo trees, semantic search by meaning (by haskellforall), advanced refactoring, blast radius, advanced file context trees, restore points... i can keep going on semantic code search and context trees are the future of agentic coding and this tool proves it the feature i loved the most is semantic search and how it gets things done 2x faster with least possible tokens it makes an agent that actually knows what it’s doing and not just guessing, it makes meaning from your code similar to RAG. if you aren't optimizing your context, you are just burning money the developer says this tool is still under development, it can have unexpected behavior and the docs need updates but the video shows the reality of how fast it can be github: get here:

forloop

226,054 Aufrufe • vor 5 Monaten

I got to try Grok 4.5 in early access in Cursor for the past few days and I absolutely enjoyed it. It feels like Opus 4.8 at 2x the speed at a much cheaper price point. I tasked it to brainstorm > plan > implement a big feature for my game (this act 1 boss fight) and it did not disappoint. - It is much smarter than Composer 2.5, during planning mode, it is able to think through my request more robustly, ensuring that edge cases are covered and makes sure to ask the right questions to confirm with me first. - It is much better at brainstorming ideas/suggestions, similar to Opus 4.8, though I think Fable still edges out a little when it comes to brainstorming ideas and suggestions - It is FAST. probably the fastest of all frontier models (Opus 4.8, GPT 5.5 etc), which makes it a joy to build with, because I can stay in the flow - It has much improved visual/animation capabilities than Composer 2.5, it can code up animations (i wanted an explosion animation with particle effects) with much, much better visuals, animation movement and timing. This is a big leap and I was so happy to see this improvement. - The best part for me is that I can just use the same model from planning down to execution without switching to a lower cost model because the price point is cheaper than other frontier models. I'll be testing this model with more challenging tasks in the next few days but I think this is going to be my main driver for vibe coding for a while. Also, its nice to see Grok back in the race. 🙌
0:34

I got to try Grok 4.5 in early access in Cursor for the past few days and I absolutely enjoyed it. It feels like Opus 4.8 at 2x the speed at a much cheaper price point. I tasked it to brainstorm > plan > implement a big feature for my game (this act 1 boss fight) and it did not disappoint. - It is much smarter than Composer 2.5, during planning mode, it is able to think through my request more robustly, ensuring that edge cases are covered and makes sure to ask the right questions to confirm with me first. - It is much better at brainstorming ideas/suggestions, similar to Opus 4.8, though I think Fable still edges out a little when it comes to brainstorming ideas and suggestions - It is FAST. probably the fastest of all frontier models (Opus 4.8, GPT 5.5 etc), which makes it a joy to build with, because I can stay in the flow - It has much improved visual/animation capabilities than Composer 2.5, it can code up animations (i wanted an explosion animation with particle effects) with much, much better visuals, animation movement and timing. This is a big leap and I was so happy to see this improvement. - The best part for me is that I can just use the same model from planning down to execution without switching to a lower cost model because the price point is cheaper than other frontier models. I'll be testing this model with more challenging tasks in the next few days but I think this is going to be my main driver for vibe coding for a while. Also, its nice to see Grok back in the race. 🙌

Danny Limanseta

1,410,316 Aufrufe • vor 21 Tagen

Cerebras inference is very fast. So fast that it changes how we think about configuring our LLMs for voice agent use cases. Kimi K2.6 is a 1T parameter reasoning model that Cerebras serves at 650 - 1,000 tokens per second (end-to-end throughput), with time to first token metrics as low as 150ms (latency). These numbers are two to three times faster than other similarly capable models. The biggest lever we get from this kind of speed is that we can use the model in reasoning mode, and still have excellent "time to first non-thinking token." This solves a big pain point we have in 2026 for voice agent use cases. Almost all recent innovation in post-training has focused on making models good at reasoning ("test time compute"). This is great, but it makes the user-facing model latency much, much slower. Which is a problem for conversational voice agents. We can run Kimi K2.6 with reasoning turned on, and get responses faster than other models produce with reasoning disabled. On my 30-turn voice agent benchmark, Kimi K2.6 with reasoning enabled ties GPT 5.1 and Haiku 4.5 with reasoning disabled, and is still about 200ms seconds faster! On my primary task agent benchmark, Kimi K2.6 is now the #2 model. It ranks just behind Gemini 3.5 Flash in "high" reasoning mode, and tied with GLM 5, Sonnet 4.6, and GPT 5.4 with reasoning set to "low." But Kimi K2.6 completes each turn in the agent loop in under 500ms. The other four models are all at least 3x slower. (Models only qualify for this benchmark if they can complete task turns at a P50 <4s.) A couple of other things that this speed buys us, for production voice agents: - Tool calls happen fast enough that we don't have to work around tool call latency in our pipeline design. - We can prompt the model to output structured data at the beginning of a response, followed by plain text for voice generation. This opens up possibilities like asking the model to do complex classification/generation tasks that influence the rest of the pipeline. For example, the model could create a detailed style prompt for a steerable TTS model, for each individual conversation turn. And, of course, you can use Kimi K2.6 with reasoning turned off. Cerebras calls this "instant" mode. Here's a video of a Cerebras Kimi K2.6 voice agent with voice-to-voice response time, measured at the client, under 500ms. This is the true response latency as perceived by the user, including all network and audio codec overhead, transcription and turn detection, Kimi K2.6 token generation, and voice generation. 500ms is, effectively, instant. So the Cerebras naming for this mode is a propos. :-)
0:40

Cerebras inference is very fast. So fast that it changes how we think about configuring our LLMs for voice agent use cases. Kimi K2.6 is a 1T parameter reasoning model that Cerebras serves at 650 - 1,000 tokens per second (end-to-end throughput), with time to first token metrics as low as 150ms (latency). These numbers are two to three times faster than other similarly capable models. The biggest lever we get from this kind of speed is that we can use the model in reasoning mode, and still have excellent "time to first non-thinking token." This solves a big pain point we have in 2026 for voice agent use cases. Almost all recent innovation in post-training has focused on making models good at reasoning ("test time compute"). This is great, but it makes the user-facing model latency much, much slower. Which is a problem for conversational voice agents. We can run Kimi K2.6 with reasoning turned on, and get responses faster than other models produce with reasoning disabled. On my 30-turn voice agent benchmark, Kimi K2.6 with reasoning enabled ties GPT 5.1 and Haiku 4.5 with reasoning disabled, and is still about 200ms seconds faster! On my primary task agent benchmark, Kimi K2.6 is now the #2 model. It ranks just behind Gemini 3.5 Flash in "high" reasoning mode, and tied with GLM 5, Sonnet 4.6, and GPT 5.4 with reasoning set to "low." But Kimi K2.6 completes each turn in the agent loop in under 500ms. The other four models are all at least 3x slower. (Models only qualify for this benchmark if they can complete task turns at a P50 <4s.) A couple of other things that this speed buys us, for production voice agents: - Tool calls happen fast enough that we don't have to work around tool call latency in our pipeline design. - We can prompt the model to output structured data at the beginning of a response, followed by plain text for voice generation. This opens up possibilities like asking the model to do complex classification/generation tasks that influence the rest of the pipeline. For example, the model could create a detailed style prompt for a steerable TTS model, for each individual conversation turn. And, of course, you can use Kimi K2.6 with reasoning turned off. Cerebras calls this "instant" mode. Here's a video of a Cerebras Kimi K2.6 voice agent with voice-to-voice response time, measured at the client, under 500ms. This is the true response latency as perceived by the user, including all network and audio codec overhead, transcription and turn detection, Kimi K2.6 token generation, and voice generation. 500ms is, effectively, instant. So the Cerebras naming for this mode is a propos. :-)

kwindla

40,319 Aufrufe • vor 2 Monaten

GROK PLAY: WHERE YOUR AI OWL LEARNS TO WIN Grok Play is xAI’s most electrifying glimpse into the future of interactive AI entertainment. Built during xAI’s 24 hour hackathon on December 6–7, 2025, by Nathan Cloos and Simon Lee, the project won first place in its track and quickly became the talk of the event. At its core, Grok Play turns your “Grok Owl,” a personalized, customizable AI avatar powered by Grok’s reasoning models, into a gamer’s alter ego. You can jump into fast-paced browser arenas like pinball shooters or futuristic Pong battles, facing friends, their Owls, or even clones of your own creation. The Owls can play autonomously, using Grok’s real-time logic to adapt tactics mid-match, or you can take control yourself and guide them to victory. Players can team up, go head-to-head, or train solo to climb local and global leaderboards. You can even build your own “Gyms,” custom arenas shaped with a Grok-assisted no-code editor where a simple command like “add gravity flips” instantly transforms gameplay. Each battle fuels Grok’s learning loop, generating gameplay data that sharpens the models behind it. Owls evolve with every move, learning your playstyle and improving over time. Source: Kathy Read , TestingCatalog News 🗞
0:46

GROK PLAY: WHERE YOUR AI OWL LEARNS TO WIN Grok Play is xAI’s most electrifying glimpse into the future of interactive AI entertainment. Built during xAI’s 24 hour hackathon on December 6–7, 2025, by Nathan Cloos and Simon Lee, the project won first place in its track and quickly became the talk of the event. At its core, Grok Play turns your “Grok Owl,” a personalized, customizable AI avatar powered by Grok’s reasoning models, into a gamer’s alter ego. You can jump into fast-paced browser arenas like pinball shooters or futuristic Pong battles, facing friends, their Owls, or even clones of your own creation. The Owls can play autonomously, using Grok’s real-time logic to adapt tactics mid-match, or you can take control yourself and guide them to victory. Players can team up, go head-to-head, or train solo to climb local and global leaderboards. You can even build your own “Gyms,” custom arenas shaped with a Grok-assisted no-code editor where a simple command like “add gravity flips” instantly transforms gameplay. Each battle fuels Grok’s learning loop, generating gameplay data that sharpens the models behind it. Owls evolve with every move, learning your playstyle and improving over time. Source: Kathy Read , TestingCatalog News 🗞

Mario Nawfal

17,572 Aufrufe • vor 7 Monaten

Nothing beats open-source! We now have a fully open-source coding agent that can compete with any proprietary solution out there: • 500+ available models • It doesn't train on your code • It doesn't charge a penny for tokens • It's open. You can change if you want. Kilo Code is the #1 app on OpenRouter right now, and the community is releasing new features on GitHub at rocket speed (about 20 releases in the last 30 days). I installed it in Visual Studio Code, but you can install it in pretty much any IDE. I need to spend more time with it, but here is a very interesting design choice: Most coding agents support two modes: an "agentic" mode for writing code and an "ask" mode for talking to the model without making any changes. Kilo Code supports 5 different modes: 1. Architect mode - Research, design, and planning 2. Debug mode - Problem solving and troubleshooting 3. Code mode - To write code. 4. Ask mode - To talk to the model. 5. Orchestrator mode - Combine every agent to solve problems. On paper, I really like this. I'll need more time to try out the specialized agents and see whether niching becomes an advantage. Here is the link to the website: Thanks to the Kilo for partnering with me on this post.
6:17

Nothing beats open-source! We now have a fully open-source coding agent that can compete with any proprietary solution out there: • 500+ available models • It doesn't train on your code • It doesn't charge a penny for tokens • It's open. You can change if you want. Kilo Code is the #1 app on OpenRouter right now, and the community is releasing new features on GitHub at rocket speed (about 20 releases in the last 30 days). I installed it in Visual Studio Code, but you can install it in pretty much any IDE. I need to spend more time with it, but here is a very interesting design choice: Most coding agents support two modes: an "agentic" mode for writing code and an "ask" mode for talking to the model without making any changes. Kilo Code supports 5 different modes: 1. Architect mode - Research, design, and planning 2. Debug mode - Problem solving and troubleshooting 3. Code mode - To write code. 4. Ask mode - To talk to the model. 5. Orchestrator mode - Combine every agent to solve problems. On paper, I really like this. I'll need more time to try out the specialized agents and see whether niching becomes an advantage. Here is the link to the website: Thanks to the Kilo for partnering with me on this post.

Santiago

73,062 Aufrufe • vor 8 Monaten

Seedance 2.0 vs. Seedance 2.0 Mini Actually, I tested this directly in Dreamina. It's already available on my account, but it seems Dreamina hasn't announced it on their social media yet 😆 Anyway, here are my first impressions. Basically, it's not bad... not bad at all. However, when I tested it with something more action-oriented, the difference was huge. It depends on your goal, but I don't think I'll use Seedance 2.0 Mini for action-heavy prompts. I could be wrong though, and I need to do more testing. For something that looks simpler and more natural, Seedance 2.0 Mini is surprisingly good. You can still notice some glitches or artifacts from time to time, but overall it's pretty decent. It's definitely not worse than the current version of Grok Imagine Video 😂😂In fact, this is what Grok Imagine should be. They're almost there but Instead, they're going backwards. I think only Tetsuo and Freeze keep saying Grok Imagine is the best model in the world LMAO 🤣🤣🤣 Still, I think Seedance 2.0 Mini delivers exactly what it promises. It's cheaper, faster, although it comes with a slight sacrifice in quality (well obviously it's a mini version of it), although not as much as I expected based on my initial tests. I think in the end, you need to identify your needs and decide when to use Mini, when to use Fast, and when to use the full version. At this stage, I can't say much yet, but I'll keep exploring it further. Meanwhile, check out the video comparison I made here.
2:04

Seedance 2.0 vs. Seedance 2.0 Mini Actually, I tested this directly in Dreamina. It's already available on my account, but it seems Dreamina hasn't announced it on their social media yet 😆 Anyway, here are my first impressions. Basically, it's not bad... not bad at all. However, when I tested it with something more action-oriented, the difference was huge. It depends on your goal, but I don't think I'll use Seedance 2.0 Mini for action-heavy prompts. I could be wrong though, and I need to do more testing. For something that looks simpler and more natural, Seedance 2.0 Mini is surprisingly good. You can still notice some glitches or artifacts from time to time, but overall it's pretty decent. It's definitely not worse than the current version of Grok Imagine Video 😂😂In fact, this is what Grok Imagine should be. They're almost there but Instead, they're going backwards. I think only Tetsuo and Freeze keep saying Grok Imagine is the best model in the world LMAO 🤣🤣🤣 Still, I think Seedance 2.0 Mini delivers exactly what it promises. It's cheaper, faster, although it comes with a slight sacrifice in quality (well obviously it's a mini version of it), although not as much as I expected based on my initial tests. I think in the end, you need to identify your needs and decide when to use Mini, when to use Fast, and when to use the full version. At this stage, I can't say much yet, but I'll keep exploring it further. Meanwhile, check out the video comparison I made here.

DAN · MXVDXN

30,516 Aufrufe • vor 1 Monat

Let's build a Deepseek Wrapper (Mobile App) without writing a single line of code! Everyone is talking about building web apps with AI, but no one is talking about building mobile apps. This video tutorial and @senior_swc template (built by ) will help you get starting building an app on iOS. This is react native so it's also compatible with Android and Web (But we will cover this in a future video). The point of this video is to get some experience in the simulator and to start building an app. In a video, soon, we will take the app we've built, test it on our phone, and put it on Test-Flight! (Without writing ANY code!) Just a warning: The DeepSeek API has been tough to rely on. So I decided to add Gemini into my app as a backup, which you can do as well (with any AI model) The code for template is on the next tweet. Timestamps: 00:00 Introduction 01:35 Setting Up Xcode and iPhone Simulator 02:29 Downloading and Configuring Cursor 04:10 Forking the Template and Initial Setup 06:10 Implementing Authentication with Clerk 08:23 Adding DeepSeek API and Testing 13:35 Enhancing the App with SupaData API 19:00 Using the Serp API to fetch video links 23:00 Figuring things out 27:30 Woohoo!!! Perplexitube Shout out to Galaxies-dev on GitHub, who's repo (chatgpt-clone-react-native), we very helpful in building this template quickly.
28:54

Let's build a Deepseek Wrapper (Mobile App) without writing a single line of code! Everyone is talking about building web apps with AI, but no one is talking about building mobile apps. This video tutorial and @senior_swc template (built by ) will help you get starting building an app on iOS. This is react native so it's also compatible with Android and Web (But we will cover this in a future video). The point of this video is to get some experience in the simulator and to start building an app. In a video, soon, we will take the app we've built, test it on our phone, and put it on Test-Flight! (Without writing ANY code!) Just a warning: The DeepSeek API has been tough to rely on. So I decided to add Gemini into my app as a backup, which you can do as well (with any AI model) The code for template is on the next tweet. Timestamps: 00:00 Introduction 01:35 Setting Up Xcode and iPhone Simulator 02:29 Downloading and Configuring Cursor 04:10 Forking the Template and Initial Setup 06:10 Implementing Authentication with Clerk 08:23 Adding DeepSeek API and Testing 13:35 Enhancing the App with SupaData API 19:00 Using the Serp API to fetch video links 23:00 Figuring things out 27:30 Woohoo!!! Perplexitube Shout out to Galaxies-dev on GitHub, who's repo (chatgpt-clone-react-native), we very helpful in building this template quickly.

Riley Brown

128,580 Aufrufe • vor 1 Jahr