Isaac

Amongst my friends, Spotify is the lowest quality consumer app we still pay for. It certainly hasnt gotten noticeably better in the last couple years (arguably worse). So, this is not the positive look Ant and Spotify are spinning here. Bigger picture, this is the problem with a lot of AI reporting. It reports completely meaningless metrics like deploys per day or LoC. Why don’t we start reporting consumer satisfaction reports? Actually end state research results. All the no nuance AI people always come out and think that this is anti AI. Again, I think AI is great and Claude is great. But this is bad marketing and makes both look like clowns.

should be obvious by now, but OpenAI and Anthropic are just gonna keep cannibalizing all their biggest customers. it’s simply too profitable for them to resist. and it’s already happening: 1. Figma partnered with Anthropic on AI design tools. then Anthropic’s product chief quit Figma’s board, and 3 days later Anthropic launched Claude Design to compete with Figma. CEO Dylan Field said Anthropic was “not consistently candid.” 2. Novo Nordisk uses Claude to help develop drugs. now Anthropic is developing drugs of its own. 3. Microsoft poured billions into OpenAI. now OpenAI is building a Jobs Platform to compete with LinkedIn, and reportedly a code repository to compete with GitHub. Microsoft owns both. 4. Harvey uses Claude to sell AI contract analysis, due diligence, and litigation tools. now Anthropic sells those same workflows through Claude for Legal. 5. Intercom used OpenAI’s Realtime API to build Fin Voice. now OpenAI sells its own voice-and-chat support agent through Presence. 6. Abridge and Ambience build clinical documentation products on OpenAI. now OpenAI sells ChatGPT for Healthcare directly to hospitals with clinical documentation built in. 7. Benchling uses Claude to power its biotech R&D platform. now Anthropic sells its own scientific workbench through Claude Science. the frontier lab playbook is simple: 1. sell their models to the world’s most valuable businesses 2. help wire them into those companies’ most valuable and sensitive work 3. map the business from the inside and find where AI can take over 4. turn those capabilities into their own products and become the customer’s competitor

Anthropic pays $785K/year for Forward Deployed Engineers In this 20-minute talk, an Anthropic FDE engineer explained the full roadmap: • step 1 → you're not selling software or time. you're selling an outcome. "a leader in oil and gas cares about throughput, not how the data is organized" • step 2 → FDE exists for one case only: technical product + non-technical buyer. everything else runs on DevRel or sales • step 3 → every AI platform is now agentic & customizable - every AI company is in that case now. "your customers have no idea what you actually do" • step 4 → the result: Palantir's ACV is $4M per customer. ServiceNow $1.2M. Workday $600K. no other public SaaS cracks $500K watch, then read the guide below ↓

SAM ALTMAN: “We are close to creating a genie that can grant any wish.”

Software quality now depends on the constraints you set around your agents. When humans manually wrote most of the code we could look at the code itself for signs of quality. Is it clean? Is it thoughtful? Is it fast? Can another engineer understand it? Does it have tests? Agents can now generate more code than people can read. When code generation scales beyond review, quality - checks for one or more of correctness, maintainability, security, performance etc - increasingly has to live somewhere else. It moves into the harness, environment and operating system around the agent. This can be the tests and deterministic checks that decide what the system is allowed to do (amongst others). Your constraints are what may eventually enable loops of agents to deliver production software reliably. They can include unit tests, property tests, acceptance tests, mutation testing and quality metrics. This back-pressure lets the system resist bad work before it becomes somebody elses problem. Set your constraints. They decide whether the code your agents generate is good enough to ship.

It’s well within Anthropic’s rights to compete in any market they choose. What’s funny, in this instance, are the number of Pharma companies, who through their unchecked use of Anthropic, are driving revenues into what they think is a model provider but is in fact a competitor lurking in the shadows thereby accelerating their own demise. I suspect any end market with reasonable ROCE that could be AI accelerated is on the table. If I were them, I’d probably do the same.

Had to reject a candidate coz she couldn't answer this question. I asked: We've an agent with access to 50+ tools, how'd u make them available to the LLM? She replied: I'll pass all of them to LLM through prompt injection & let it decide. I countered: Wouldn't it cause context bloat? Shouldn't you narrow the needed tools list based on the prompt or usage before sending? No answer. What would you've answered to get selected?

Is Bradley Barcola the player to replace Mo Salah at Liverpool? 🤔🇫🇷

Work mode 🔛

The AI productivity gap between solopreneurs and enterprises is going to be fascinating A solo developer can adopt a new workflow overnight An enterprise has security reviews, compliance, legacy systems, approval chains, and hundreds of developers to coordinate Same models Completely different speed limits

Mo Salah averaged 5010mins a season He played 5216 in 25/26 Barcola averages 2704 but did manage 3592 last season including World Cup Gakpo is also a minutes monster, averaging 4355mins a season. It’s easy to say sell him but somebody needs to actually play Rio really shouldn’t be playing >2000mins at this point in his career and he only played 1278 last season Who is actually filling the mins

One player Liverpool should never have let leave is Fabinho. He was only 29 (now 32). Liverpool need cover for Right back, Center back and need a true DM. Fabinho is able to play at Right back, Center back and is a Defensive midfielder Should Liverpool look at him? Any other currently still playing players that you wish Liverpool hadn’t moved on?

Genuinely the first time I’ve ever seen a football club , in these touch exercises, actually serve footballs that aren’t just the same distance, same weight on the floor and serve some unpredictable variance. Technical practices are some of the worst at clubs. It’s so routine and formulaic based with no adaptability or random variance . Even though football is full of random variance all the time and you should always be training that. When do you ever see the club for example, randomly bounce the awkward pass in, make a player deal with it? Pretty much never even though that stuff happens all the time in games.

Zero alcohol is better than any amount. Andrew Huberman’s practical ceiling for people who aren’t dealing with alcohol use disorder: two drinks per week. Beyond that, data from the UK Brain Bank shows a measurable pattern, every extra drink per week is linked to a bit more brain atrophy and neuron loss. Large-scale imaging studies (including UK Biobank analyses) have found a dose-dependent association between higher weekly alcohol intake and reduced gray-matter volume, even at levels once considered moderate. Clear dose-response. The less, the better for the brain.

This sounds true if you are highly expeienced and proficient at systems level programming, like George Hotz or Jonathan Blow. Most people are not. For most people, the limiting factor *is* the code. Not typing it, but figuring out what to type in the first place. You may know the theoretical solution to a problem, but getting it to actually happen requires solving many sub-problems that most people are not capable of, either because they lack the mental capacity to handle it, or for lack of practice. Managing complexity only comes to play when you actually have a system that is large enough where complexity actually becomes an issue. Most programmers never reach that level, yet LARP into "managing complexity" with their towers of abstractions based "architecture". The sub-tweet is obviously about AI progrmaming, so I will give my take on that: * AI is not for writing the code you already know how to write; it's for writing code you _don't know_ how to write. * You can still manage complexity without reviewing every single line of code: just manage system architecture and boundaries between sub-systems. It's not easy per se, and, to concede a point, you probably cannot acquire the "taste" for where to draw boundaries without having spent a long amount of time hand coding a semi-complex system from scratch by yourself.