So could I in practice train it on all my psychology books, materials, reports, case study and research papers and then run it on demand on a 1xH100 node - https://getdeploying.com/reference/cloud-gpu/nvidia-h100 whenever I have a specialised question?
I focus more on reading code & prompting claude to write code for me at a high level. I also experiment a lot. I don't write code anymore by hand except in very rare cases. I ask claude for questions about the code to build understanding. I have it produce documentation, which is then consumed into other prompts. Often, claude code will need several minutes on a task so I start another task. My coding throughput on a day to day basis is now the equivalent of about 2-3 people.
I also use gemini to try out trading ideas. For example, the other day I had gemini process google's latest quarterly report to create a market value given the total sum of all it's businesses. It valued google at $215. Then I bought long call options on google. Literally vibe day trading.
I use chat gpt sora to experiment with art. I've always been fascinated with frank lloyd wright and o4 has gotten good enough to not munge the squares around in the coonley playhouse image so that's been a lot of fun to mess with.
I use cheaper models & rag to automate categorizing of my transactions in Tiller. Claude code does the devops/python scripting to set up anything google cloud related so I can connect directly to my budget spreadsheet in google sheets. Then I use llama via openrouter + a complex RAG system to analyze my historical credit card data & come up with accurate categorizations for new transactions.
This is only scratching the surface. I now use claude for devops, frontend, backend, fixing issues with embedder models in huggingface candle. The list is endless.
I'm a CTO who makes purchasing decisions. There are numerous products I likely would have purchased, but I either find a substitute or just go without because I won't play the stupid "let's get on a call" game.
If your website doesn't give me enough information to:
1. Know enough about your product to know that it will (generally speaking) meet my needs/requirements.
2. Know that the pricing is within the ballpark of reasonable given what your product does.
Then I will move on (unless I'm really desparate, which I assure you is rarely the case). I've rolled-my-own solution more than once as well when there were no other good competitors.
That's not to say that calls never work or don't have a place, because they definitely do. The key to using the call successfully (with me at least) is to use the call to get into true details about my needs, after I know that you're at least in the ballpark. Additionally, the call should be done efficiently. We don't need a 15 minute introduction and overview about you. We don't need a bunch of small talk about weather or sports. 2 minutes of that is ok, or when waiting for additional people to join the call, but beyond that I have things to do.
I know what my needs are. I understand you need some context on my company and needs in order to push useful information forward, and I also understand that many potential customers will not take the lead in asking questions and providing that context, but the sooner you take the temperature and adjust, the better. Also, you can get pretty far as a salesperson if you just spend 5 minutes looking at our website before the call! Then you don't have to ask basic questions about what we do. If you're willing to invest in the time to get on a call, then it's worth a few minutes of time before-hand to look at our website.
I wouldn't go as far to say "everything costs less" but it is pretty well that established that poverty is very expensive.
A couple of key examples:
Food deserts often mean that groceries are more expensive in poorer areas as opposed to neighboring rich ones. Additionally, bulk food is cheaper but requires having enough funds to buy more than your immediate needs.
It is generally cheaper to own your own home than to rent and low income people are going to pay higher interest on the same home loan.
It is always cheaper for rich people to borrow money than poor people and poor people are often forced into debt in situations where rich people can dip into savings. Having to pay interest on your rainy day debt is way more expensive than getting paid interest on your rainy day savings.
That last one is huge, and tends to compound across all kinds of other areas, increasing the effective price that poor people pay for almost everything.
In the most general sense, it is often feasible to spend more money up front to save money down the road. The amount of interest poor people have to pay to do this reduces or even totally wipes out any savings.
This is all pretty well documented and studied. It's part of the unfortunate feedback cycle at the bottom of the economic bracket that makes climbing back out harder the poorer you get.
I called myself a Semiotics Engineer for 4 years, but the title didn't catch. I did domain analysis, logical model creation, concrete model creation in XML/OWL/KML, model review and improvement, semantic reasoning-based system design/implementation, and message system design/implementation. This was before the rise of ML.
Code traces are metrics. Run times per function calls metrics, count of specific function call metrics.
Otel is an attempt to package such arithmetic.
Web apps have added so many layers of syntax sugar and semantic wank, we’ve lost sight its all just the same old math operations relative to different math objects. Sets are not triangles but both are tested, quantified, and compared with the same old mathematical ops we learn by middle school.
Another Obsidian alternative which I use every day is Anytype[1]. It's fully open source however under their own license which has some interesting terms to discourage commercial adoption. They seem to be very focused on individual use. The user experience is similar to Notion with some subtle differences, but overall very positive. The biggest plus for me was offline p2p sync and a really solid mobile app.
I have been speculating on Nvidia stock. It has been following an exponential curve. I have curve fitted the stock price, and determined that:
- in 2027, the market cap of Nvidia will be approximately $10 trillion - approximately half of US GDP
- in 2035, the market cap will be $1,500 trillion. Far greater than the total amount of value produced throughout human history.
It would be very unwise to bet against me; up to this point, my curve fit has been completely correct. Naysayers have been squashed. By my calculations, which are based on the laws of mathematics and completely indisputable -- the total market cap of Nvidia will be well over 10 quadrillion dollars by 2040.
I'm actually growing concerned. In our lifetime, Nvidia stock might become so uncountably valuable that even an attempt to calculate its value will consume most of the energy available in the known universe. Because I'm serious about my math, I've put some confidence intervals on my estimate; but, dear reader, even that doesn't save us. In the conservative case, it only forestalls the inevitable by just a few years.
I suggest we assemble a task force to ponder the inevitable problems that will arise from this set of facts. Humbly, I will submit myself as the leader of such a task force. Because of the gravity of the situation, I think a few million dollars, a small team, and perhaps a few hundred or thousand shares of Nvidia -- will help to fully understand the implications of this dire situation.
> Apache Ballista and Polars do Apache Arrow and SIMD.
> The Polars homepage links to the "Database-like ops benchmark" of {Polars, data.table, DataFrames.jl, ClickHouse, cuDF, spark, (py)datatable, dplyr, pandas, dask, Arrow, DuckDB, Modin,} but not yet PostgresML? https://h2oai.github.io/db-benchmark/ *
> ElectricSQL is a local-first software platform that makes it easy to develop high-quality, modern apps with instant reactivity, realtime multi-user collaboration and conflict-free offline support.
> Local-first is a new development paradigm where your app code talks directly to an embedded local database and data syncs in the background via active-active database replication. Because the app code talks directly to a local database, apps feel instant. Because data syncs in the background via active-active replication it naturally supports multi-user collaboration and conflict-free offline
> The problem is 100% the use of LLMs to pull the content in an unsanctioned manner
If the public facing web wasn't crawlable Google and many other things wouldn't be possible. What are you saying that YouTube should not be viewable unless someone is properly authed? Take it up with YouTube - they could require logins to view videos if they wanted but it would be a worse product.
When a user posts something on YouTube and checks "Allow embeds" they have not only given their video to YouTube but are totally cool with people sharing it around. Who are you - their lawyer? Even most YouTube creators do not share your opinion as I can see most are allowing embeds. The point of mentioning embeds is it's both:
1) Credit to the creator, including the revenue share which you were incorrect when you said it lowers their revenues - they still get the ad revenue from an embed
2) An indicator that the creator wants their content to be shared, since they have the option to disable them if they want and choose not to
The adjacent point you seem to be making is that nobody should be allowed to crawl information and present it as their own. But that's what a lot of the internet is. It's what a search engine is, it's the source of most online encyclopedias, news sites, and so-on.
It just doesn't logically follow that if somebody uploads a video to YouTube, that nobody is allowed to summarize it in text form. That's a very normal thing that is done online.
With that, you can replace a Dialer in Go that connects sockets, effectively wrapping sockets with Wireguard. Since it does that in userspace, you get no tun/tap. This is all open-sourced by @dpeckett
I had a socks server running in docker that turned Forticlient, the worst corporate vpn solution ever, into socks so then I could use access the internal urls in one container, still have fast internet elsewhere, and not expose my computer's entire network traffic to Forticrap.
Automatization has made it possible to do all these stuff at such a scale that Google can spy on everyone all the time.
Stasi was limited by that the whole of DDR can't work at Stasi.
Instead of the government joining in on the fun, maybe it would be good to e.g. close down Google (split up the spy and search parts, which essentially is closing down Google since they have relatively nothing without the spying).
IMO you lose so much more by sacrificing spontaneous conversation & ideation that results. You also lose the ease of just walking over to someone to ask a question. You also lose an unbelievable amount for anyone who lacks experience - training is AWFUL remote. Not even close.
It's not perfect but a group of aligned people in the same physical working space will just dominate a similar group spread apart that has to use chats & zoom to communicate. Management has got to be seeing this, in various forms, across multiple business segments.
Some things I'd recommend, off the top of my head:
- Do not grab phone/computer etc. and mindlessly browse first thing in the morning. (Or before bed. Or at any time really.) But doing it first thing really starts your day on the wrong foot.
- When seeking to relax, do not mindlessly browse the internet/social-media/tv. Read an enjoyable book. This is an order of magnitude more fulfilling and beneficial to you. And genuinely more relaxing: screens are stimulating, and might let you 'relax' in the sense that you can momentarily be completely absorbed in something 'other', and forget your day to day life; but they don't relax you in the sense of being calm and contemplative (in general, in my experience).
- Reduce instant gratification from as many areas as possible. Do things that are rewarding longer term. Like reading, cooking, growing plants, hiking, etc.
- Cut video games.
- Block facebook + reddit + sites you waste a lot of time on, from main computer. Maybe have a secondary device you use to access these sites, for a set period each day (I recommend this mainly because it can be quite difficult to maintain a social life without facebook, (which is a terrible state of affairs)). Have days where you don't go onto these sites at all.
- Spend as little time on screens as is possible
-> if you can work on paper do so
- have a regular exercise regime. eg. swim/run. Doing first thing in the morning really helps set your day on the right track, you have already exerted a good amount of self discipline, and achieved something, and this makes it easier to continue being disciplined.
