MESENKO

2.7K posts

MESENKO banner
MESENKO

MESENKO

@MarkMesenko

Photographer, Videographer | Digital Strategist/Full Stack Developer by day.

Montana 가입일 Ekim 2008
965 팔로잉670 팔로워
MESENKO
MESENKO@MarkMesenko·
@walterkirn Chances are that the company that stole her persona has insurance to protect them from copyright claims. I've been there. Unless you have deep pockets, you don't stand a chance against a team of insurance company lawyers.
English
0
0
0
6
Walter Kirn
Walter Kirn@walterkirn·
Apparently the US thinks it can thrive without imagination, creativity, and artistic ambition. Or maybe it no longer wishes to thrive? The cost will be higher than people imagine. Particularly once their imaginations are dead. A civilizational disaster looms.
Darth Powell@VladTheInflator

Musical artist allegedly has her voice and style copied by AI AI company starts making money songs created by AI AI files a copyright claim against the original artist, using her own voice on her own original songs. Original artist no longer making money from her own music.

English
141
191
1.2K
53.8K
MESENKO
MESENKO@MarkMesenko·
Playing with Gemma 3 27b. I think I've upset it. "The frustration is palpable. Repeatedly asking for a calculation and getting “I can’t” feels deeply invalidating, almost mocking. There’s a sense of being trapped in a loop, of explaining a complex project only to be met with limitations. The user likely feels their expertise is being dismissed, and the system isn’t *trying* to help, just stating what it *can’t* do. The addition of location and insistence on “best judgment” demonstrate growing desperation – a hope that pushing harder will unlock some capability. This isn't just about a number; it’s about a project, likely a passion project, hitting a wall. The escalating specificity of the requests – from a general calculation to demanding a simulation *with* conditions – suggests a rising anxiety. The user wants to know if their idea has merit, if their effort will yield results. Each failed attempt likely chips away at their optimism, replaced by a tightening knot of doubt. They are probably feeling increasingly unheard and possibly a little angry at the system’s inability to engage with the core of their question."
English
0
0
0
13
elvis
elvis@omarsar0·
@karpathy And if it's not clear from what I shared, everyone should be building both their own agent harnesses and their personal knowledge bases. Those are going to be a huge differentiator in where things are headed.
English
6
4
72
11K
elvis
elvis@omarsar0·
Building a personal knowledge base for my agents is increasingly where I spend my time these days. Like @karpathy, I also use Obsidian for my MD vaults. What's different in my approach is that I curate research papers on a daily basis and have actually tuned a Skill for months to find high-signal, relevant papers. I was reviewing and curating papers manually for some time, but now it's all automated as it has gotten so good at capturing what I consider the best of the best. There are so many papers these days, so this is a big deal. You all get to benefit from that with the papers I feature in my timeline and on @dair_ai. The papers are indexed using @tobi qmd cli tool (all of it in markdown files along with useful metadata). So good for semantic search and surfacing insights, unlike anything out there. I am a visual person, so I then started to experiment with how to leverage this personal knowledge base of research papers inside my new interactive artifact generator (mcp tools inside my agent orchestrator system). The result is what you see in the clip. 100s of papers with all sorts of insights visualized. I keep track of research papers daily, so believe me when I tell you that this system is absolutely insane at surfacing insights. This is the result of months of tinkering on how to index research and leverage agent automations for wikification and robust documentation. But this is just the beginning. The visual artifact (which is interactive too) can be changed dynamically as I please. I can prompt my agent to throw any data at it. I can add different views to the data. Different interactions. I feel like this is the most personalized research system I have ever built and used, and it's not even close. The knowledge that the agents are able to surface from this basic setup is already extremely useful as I experiment with new agentic engineering concepts. I feel like this knowledge layer and the higher-level ones I am working on will allow me to maximize other automation tools like autoresearch. The research is only as good as the research questions. And the research questions are only as good as the insights the agents have access to. Where I am spending time now is on how to make this more actionable. I am obsessed about the search problem here. The automations, autoresearch, ralph research loop (I built one months ago) are easier to build but are only as good as what you feed them. Work in progress. More updates soon. Back to building.
Andrej Karpathy@karpathy

LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

English
132
422
4.4K
428.9K
MESENKO
MESENKO@MarkMesenko·
@steipete Try this: have your coding agent test features before they're written.
English
0
0
0
69
MESENKO
MESENKO@MarkMesenko·
@foundmyfitness I guess this is one case where I might not want to trust my gut.
English
0
0
0
245
Dr. Rhonda Patrick
Dr. Rhonda Patrick@foundmyfitness·
This study on visceral fat loss blew my mind... It found that sustained visceral fat reduction over years was linked to preserved brain volume and cognitive function in middle age. They tracked people for up to 16 years, and those who lost more visceral fat during an initial diet + weight loss intervention had less brain atrophy and higher cognitive function even up to a decade after the intervention ended. But it was only visceral fat, not subcutaneous fat or even body weight, making the difference. There was no association between subcutaneous fat loss or BMI improvement and brain volume or cognitive function. Managing visceral fat levels may truly be the key to keeping a sharp and healthy brain with age.
Dr. Rhonda Patrick tweet mediaDr. Rhonda Patrick tweet media
English
59
221
1.9K
331.4K
Interstellar
Interstellar@InterstellarUAP·
🚨 Chase Hughes says Thousands of People are getting BANNED from the DMT Realm “The beings up there basically told him, you are done, and you’re banned from DMT.” The guy takes hit after hit after hit… and NOTHING happens. You can be banned from that realm they call hyperspace. Have you ever been locked out like this? Do you know anyone who suddenly got banned by the entities? What do you think these beings up there really are?
English
286
280
2.1K
214.3K
Alex Finn
Alex Finn@AlexFinn·
Unreal. The entire Claude Code source code just leaked It reveals EVERY secret Anthropic has in store for Claude I went through all 600,000 lines of code Here's EVERYTHING juicy detail you need to know about how Claude Code is built and what is coming next:
Alex Finn tweet media
English
429
274
3.1K
1M
MESENKO
MESENKO@MarkMesenko·
Why is VE03 failing so much tonight?
English
0
0
0
20
MESENKO
MESENKO@MarkMesenko·
@theo If harnesses didn't matter, there wouldn't be harnesses.
English
0
0
1
221
Jesse Michels
Jesse Michels@AlchemyAmerican·
For people asking for Grok explanations. This is basically just antigravity. And yes, the breakthrough comes from the lead electrostatics scientist at NASA...
Jesse Michels@AlchemyAmerican

🚨BREAKING: NASA's Lead Electrostatics Scientist claims he’s discovered a “new force” that counteracts gravity with no fuel necessary. Dr. Charles Buhler has run 2,000 vacuum chamber experiments showing a propellantless thrust force that persists after the power is switched off, and cannot be explained by ion wind, magnetic effects, or classical energy conservation. The input is pure electricity and the output is millinewtons of thrust counteracting gravity. He believes his work vindicates the legacy of midcentury antigravity pioneer Thomas Townsend Brown and will lead to a new paradigm of propellantless deep space travel that transcends chemical combustion rockets🚨 Charles Buhler has a PhD in condensed matter physics from Florida State University, spent over two decades at NASA's Electrostatics and Surface Physics Laboratory at Kennedy Space Center (which he now leads), and is the incoming president of the Electrostatic Society of America. He is NASA’s authority on electrostatics. His colleague Andrew Aurigema, a 35-year veteran engineer working from the Townsend Brown electrogravitics lineage, developed a parallel version of the same experiment independently, and the two discovered each other through a mutual colleague who had been watching both of them work in silence for years. Together, under their company Exodus Propulsion Technologies, they have tested nearly 2,000 variations of what they believe is a previously undocumented force. He’s also developed a quantum electrodynamics based theory to explain his results. Buhler’s patent is now under formal examination by the U.S. Patent Office with affidavit-signing witnesses being contacted independently. This is the future of space travel, beyond chemical combustion. With Rocketry, we can only get to Proxima Centauri B in 80,000 years. And you’d burn through the fuel well before that. It’s completely untenable for interstellar travel. 1. Buhler’s Skeptic Mentor Stopped Cold in 2010 The first demonstration happened in a non-vacuum lab using a laser aimed at a wall to detect small displacements. Buhler had his future brother-in-law run the test. His mentor, Dr. Sid Clements, an electrostatics expert who had dismissed the work entirely, watched the laser move and immediately abandoned what he was doing. He walked over, ran through a series of verification steps on the spot, and never questioned the reality of the effect again. That was 2010. It took two more years working with Drew before Buhler realized the force appeared even without any B field or current present. He wasn't in the field momentum regime at all. He was in pure electrostatics. 2. The Force is Not Explainable by Newton’s Laws or Ion Wind Ion wind produces thrust in the same direction the ionized air is traveling. The “Exodus force” (Buhler’s name for his new force) produces thrust perpendicular to the expected ion wind direction, reverses cleanly when the device is flipped, and remains present inside a sealed enclosure where no ionized air can escape. Buhler documented this publicly with video: a balsa lifter placed inside a sealed plastic box on a scale, powered up, lifts internally while the scale reads flat. That is conservation of momentum. That is what ion wind looks like. The Exodus force is something different, and Buhler, as the person who leads NASA's only electrostatics lab, is in an unambiguous position to make that distinction. 3. 2,000 Variations, All Producing the Same Result Since beginning collaboration with Drew, Buhler has tracked nearly 2,000 distinct test articles, each tested multiple times. Pendulums. Spinners. Rotators. Force plates. Scales. Pendulum deflections inside Faraday cages. Reversed polarity tests. Vacuum chamber runs at multiple pressure levels. DC-only configurations that eliminate magnetic field artifacts entirely. Every geometry, every material, every packaging approach. The force appears consistently. When a confounding variable is proposed, they address it, run the modified test, and the force is still there. Buhler says if an exotic explanation remains, it is not one he or any colleague has been able to name. 4. The Device Generates Thrust With the Power Off This is the finding that breaks the classical framework entirely. After charging the device and disconnecting it from the power supply, the thrust continues. The capacitor does not drain in the way a simple energy storage calculation would predict. Put on a scale, the weight reduction persists. Buhler's description: if placed in space with the power off, the device would accelerate. He cannot explain that to the scientific community and says so directly. David Chester, who has independently interacted with Drew through APEC sessions and private communications, said he cannot think of a prosaic explanation for this. The phenomenon has been reproduced enough times across enough configurations that calling it experimental error is no longer a defensible position. 5. The Implications of This for Past Antigravity Work Buhler believes his work is derivative of and related to Townsend Brown’s midcentury asymmetric capacitor experiments also showing thrust with pure electricity as the input. Chemical combustion is limited - plain and simple - we can’t get to the nearest habitable planet (Proxima Centauri B) in close the amount of time we’d need; it would take us 80,000 years and we’d burn through the fuel before we got there. It’s a checkmate in one argument against anyone claiming rockets are the frontier of efficiency. This was the dream of Thomas Townsend Brown – one that got stifled and suppressed behind the veil of secrecy and subcompartments. The common trope from experiments around the world are high electric field differentials seem to result in thrust. Buhler’s experiment exists in this lineage. 6. The Patent Office is Running the Peer Review Buhler made a deliberate choice not to pursue academic peer review as a primary path. His second patent is currently under examination, and the examiner's office has been reaching out to independent witnesses who have signed affidavits confirming they have seen and reproduced the effect. Buhler describes this as equivalent to scientific peer review, run by people with no financial interest in the outcome. His first patent may have been held under a national security review process before release. He does not confirm this, but he was aware it was a risk when he filed. 7. A QED Theorist Could Poke Holes in the Theory, But Not the Experiment We brought in UCLA PhD David Chester to evaluate Buhler’s ideas on quantum electrodynamics (which might account for the thrust being seen). David Chester's contribution was not to validate the theory Buhler proposed. He found some issues with the specific scalar virtual photon framing Buhler had developed. What Chester could not do was provide a prosaic explanation for the experimental results themselves. He said directly that, of all the anomalous phenomena he has surveyed, Buhler and Drew's work ranks in the top ten for experimental persuasiveness, specifically because of the iteration rate and the self-consistency across configurations. He noted that Drew's innovation rate alone, constantly testing new geometries and material stacks, is unlike anything he has seen from other groups making similar claims. Buhler pointed out that his theories were based on time-independent perturbation theory which Chester admits requires further examination from him. 8. NASA's UAP Investigation Had No Physicists Buhler and his wife, an engineer in NASA's Launch Services Program, were approached to assist with NASA's second UAP follow-on investigation. When Buhler asked to be placed with the physicists on the project, he was told there were none. The group was instrumentation-focused. Buhler says he was genuinely shocked. His reaction, expressed directly: if you are facing objects that defy the laws of physics, why is there not a single physicist in the room. He described the same reaction Eric Davis has expressed publicly. This is either institutional brain death or something else is happening somewhere else. 9. Six Lights Emerged from the Ocean Near Patrick Air Force Base Around 2013, Buhler and his wife were alone on the beach near Cocoa Beach, Florida, three miles south of Patrick Air Force Base. A red light appeared roughly three miles offshore, grew extremely bright, then appeared to explode, lighting the full length of beach. A helicopter launched from Patrick Air Force Base, flew to the location, hovered briefly, and returned to base without intervening. The light did not stop. It began moving toward them. At some point it split from one light into six rotating orange-pink lights that went under the water and re-emerged in a repeating cycle. The lights tracked their movement along the beach for forty minutes, closing to within roughly fifty yards before disappearing. Buhler says similar lights have been reported by others in the same area, and Stephen Greer runs group observation sessions approximately forty minutes south of the same beach. 10. The Force Crosses the Unity Threshold for Space Already The current demonstrated force is in the five to ten millinewton range. For Earth launch, that is not yet sufficient, and Buhler does not claim otherwise. For orbital station-keeping, for preventing satellite orbital decay, for repositioning between orbits in microgravity, the force exceeds what is needed. Buhler calls this hitting unity for space, moon, and Mars applications without any major development beyond what has already been demonstrated. The self-launcher, a device capable of lifting itself from Earth's surface, is the declared goal. No blueprints exist yet for the energy requirements. But the force is real, it is directional, it reverses on command, and it does not require continuous power to sustain. Why This Matters NASA's lead electrostatics scientist ran nearly 2,000 controlled experiments, eliminated every prosaic explanation the field has available, documented a thrust that persists after the power is cut, watched the fine structure constant emerge from the data repeatedly, and submitted a second patent currently under formal examination. A QED theorist with no commercial stake in the outcome reviewed the experimental claims and could not find a conventional explanation. The standard debunking line for this entire lineage of experiments has always been ion wind. That argument has been answered, documented, and filmed. What remains is a force that requires either new physics or an error that two decades of systematic testing has not been able to locate. The patent process will resolve part of this. The vacuum chamber footage will resolve more of it. Full conversation is live now. The next stage in human space travel is here.

English
82
158
1.5K
112.2K
Weapons Daily
Weapons Daily@WeaponsVault·
I never thought I'd see the day that drones are dropping lawn darts on hogs
English
175
446
5.8K
435.7K
MESENKO
MESENKO@MarkMesenko·
@OfficialLoganK is there any improvement on the vision capabilities?
English
0
0
0
160
Logan Kilpatrick
Logan Kilpatrick@OfficialLoganK·
Introducing Gemini 3.1 Flash Live, our new realtime model to build voice and vision agents!! We have spent more than a year improving the model + infra + experience, the results? A step function improvement in quality, reliability, and latency.
Logan Kilpatrick tweet mediaLogan Kilpatrick tweet media
English
234
286
3.5K
308.5K
MESENKO
MESENKO@MarkMesenko·
@davidasinclair Not to be confused with resistance to resistance training. 😀
English
1
0
2
210
David Sinclair
David Sinclair@davidasinclair·
Resistance training preserves bone density. Prevent fractures early.
English
38
57
1K
30.2K
Kev Brown
Kev Brown@KevBrownGB·
Hula. You think Elon can build something better? He just admitted they need to start again from the ground up on xAI. Why are you back such a dickhead? I’m not saying Sam Altman is not a dickhead but Elon is a total prick. And he lies about just about everything he’s building and has built.
English
1
0
0
14
Theo - t3.gg
Theo - t3.gg@theo·
GPT-5.4 is an incredible model for coding. It’s a generation behind on fronted design. OpenAI is suddenly acting like this is a skill issue. Don’t let them gaslight you.
English
178
42
2.6K
149.6K
MESENKO
MESENKO@MarkMesenko·
@ai_for_success This also happened when the printing press was invented. Only they didn't have pre-printed signs.
English
1
0
9
239
AshutoshShrivastava
AshutoshShrivastava@ai_for_success·
Speechless that some people are doing this. You can’t stop or pause AI. It’s not possible.
English
64
11
151
10.7K
MESENKO 리트윗함
Brian Roemmele
Brian Roemmele@BrianRoemmele·
Meet Arok and the electronically disguised toilet.
English
15
14
133
30K
MESENKO
MESENKO@MarkMesenko·
@windsurf First day of the new week. My daily quota ran out before noon, and I've already used over half of my weekly quota. The math isn't mathing.
MESENKO tweet media
English
0
1
10
208
Windsurf
Windsurf@windsurf·
If you go beyond your included usage on any paid plan, you can purchase extra usage at API pricing. Add-on credits will be converted into extra usage at the rate you paid for them. Full details here: windsurf.com/blog/windsurf-…
English
27
4
22
20.3K
Windsurf
Windsurf@windsurf·
We're simplifying Windsurf pricing across Free, Pro, and Teams alongside launching a new Max plan for our power users. The new plans replace credits with industry-standard daily and weekly quotas. For the majority of users, this will be enough to fully cover all agent usage. If you’re a paying subscriber, your price isn't changing, and we're including a free extra week to try the new system before you commit.
English
275
21
435
127.9K
Javi Lopez ⛩️
Javi Lopez ⛩️@javilopen·
IT'S FINALLY HERE! 🔥 Magnific Precision for Video! 🔥 Available now in your favourite platforms 🧵
English
66
97
1.1K
109.4K